Patch-ID# 110539-17 Keywords: netbackup 3.4 unix server client nb_34_6 Synopsis: NetBackup 3.4 UNIX server and client fix Date: Oct/08/2003 Install Requirements: None Solaris Release: 2.6 7 8 9 SunOS Release: 5.6 5.7 5.8 5.9 Unbundled Product: Veritas NetBackup Unbundled Release: 3.4 Xref: Topic: Veritas NetBackup 3.4 jumbo patch Relevant Architectures: sparc BugId's fixed with this patch: 4324687 4356569 4387565 4397701 4406098 4429632 4432588 4499273 4500422 4503882 4507040 4523614 4641151 4641305 4652724 4666899 4678476 4700521 4700664 4759007 4836227 4927029 Changes incorporated in this version: 4927029 Patches accumulated and obsoleted by this patch: Patches which conflict with this patch: Patches required with this patch: Obsoleted by: Files included with this patch: /openv/java/allJBP.jar /openv/java/allNB.jar /openv/lib/client/Solaris/Solaris2.6/libvafs.so /openv/netbackup/bin/admincmd/bpauthsync /openv/netbackup/bin/admincmd/bpbackupdb /openv/netbackup/bin/admincmd/bpclinclude /openv/netbackup/bin/admincmd/bpcllist /openv/netbackup/bin/admincmd/bpconfig /openv/netbackup/bin/admincmd/bpdbjobs /openv/netbackup/bin/admincmd/bpduplicate /openv/netbackup/bin/admincmd/bpexpdate /openv/netbackup/bin/admincmd/bpgetconfig /openv/netbackup/bin/admincmd/bpimport /openv/netbackup/bin/admincmd/bplocaladdrs /openv/netbackup/bin/admincmd/bpmedia /openv/netbackup/bin/admincmd/bpmedialist /openv/netbackup/bin/admincmd/bpminlicense /openv/netbackup/bin/admincmd/bprecover /openv/netbackup/bin/admincmd/bpretlevel /openv/netbackup/bin/admincmd/bpsetconfig /openv/netbackup/bin/admincmd/bpverify /openv/netbackup/bin/admincmd/get_license_key /openv/netbackup/bin/admincmd/nbdbsetport /openv/netbackup/bin/admincmd/nbdbsetpw /openv/netbackup/bin/admincmd/nbdbstop /openv/netbackup/bin/bpadm /openv/netbackup/bin/bpbrm /openv/netbackup/bin/bpdbm /openv/netbackup/bin/bpdm /openv/netbackup/bin/bpinst /openv/netbackup/bin/bpjava-susvc /openv/netbackup/bin/bpps /openv/netbackup/bin/bprd /openv/netbackup/bin/bpsched /openv/netbackup/bin/bptm /openv/netbackup/bin/goodies/K77netbackup /openv/netbackup/bin/goodies/NCVU.README /openv/netbackup/bin/goodies/NCVU.tar /openv/netbackup/bin/goodies/S777netbackup /openv/netbackup/bin/goodies/S77netbackup /openv/netbackup/bin/goodies/bp.kill_all /openv/netbackup/bin/goodies/bpend_notify.oracle_bli /openv/netbackup/bin/goodies/bpstart_notify.oracle_bli /openv/netbackup/bin/goodies/ms_exchange.filelist /openv/netbackup/bin/goodies/new_clients /openv/netbackup/bin/goodies/post_checkpoint_notify.oracle_bli /openv/netbackup/bin/goodies/rc.veritas.aix /openv/netbackup/bin/goodies/xbpmon /openv/netbackup/bin/index_clients /openv/netbackup/bin/update_clients /openv/netbackup/bin/update_dbclients /openv/netbackup/client/Solaris/Solaris2.6/bp /openv/netbackup/client/Solaris/Solaris2.6/bpbackup /openv/netbackup/client/Solaris/Solaris2.6/bpbkar /openv/netbackup/client/Solaris/Solaris2.6/bpcd /openv/netbackup/client/Solaris/Solaris2.6/bpclimagelist /openv/netbackup/client/Solaris/Solaris2.6/bpclntcmd /openv/netbackup/client/Solaris/Solaris2.6/bpdynamicclient /openv/netbackup/client/Solaris/Solaris2.6/bphdb /openv/netbackup/client/Solaris/Solaris2.6/bpjava-msvc /openv/netbackup/client/Solaris/Solaris2.6/bpjava-usvc /openv/netbackup/client/Solaris/Solaris2.6/bplist /openv/netbackup/client/Solaris/Solaris2.6/bpmount /openv/netbackup/client/Solaris/Solaris2.6/bprestore /openv/netbackup/client/Solaris/Solaris2.6/client_config /openv/netbackup/client/Solaris/Solaris2.6/install_client /openv/netbackup/client/Solaris/Solaris2.6/libvopie.so /openv/netbackup/client/Solaris/Solaris2.6/tar /openv/netbackup/client/Solaris/Solaris2.6/version /openv/netbackup/client/Solaris/Solaris2.6/xbp /openv/netbackup/client/Solaris/Solaris_x86_2.6/bp /openv/netbackup/client/Solaris/Solaris_x86_2.6/bpbackup /openv/netbackup/client/Solaris/Solaris_x86_2.6/bpbkar /openv/netbackup/client/Solaris/Solaris_x86_2.6/bpcd /openv/netbackup/client/Solaris/Solaris_x86_2.6/bpclimagelist /openv/netbackup/client/Solaris/Solaris_x86_2.6/bpclntcmd /openv/netbackup/client/Solaris/Solaris_x86_2.6/bpdynamicclient /openv/netbackup/client/Solaris/Solaris_x86_2.6/bphdb /openv/netbackup/client/Solaris/Solaris_x86_2.6/bpjava-msvc /openv/netbackup/client/Solaris/Solaris_x86_2.6/bpjava-usvc /openv/netbackup/client/Solaris/Solaris_x86_2.6/bplist /openv/netbackup/client/Solaris/Solaris_x86_2.6/bpmount /openv/netbackup/client/Solaris/Solaris_x86_2.6/bprestore /openv/netbackup/client/Solaris/Solaris_x86_2.6/client_config /openv/netbackup/client/Solaris/Solaris_x86_2.6/install_client /openv/netbackup/client/Solaris/Solaris_x86_2.6/libvopie.so /openv/netbackup/client/Solaris/Solaris_x86_2.6/tar /openv/netbackup/client/Solaris/Solaris_x86_2.6/version /openv/netbackup/client/Solaris/Solaris_x86_2.6/xbp /openv/volmgr/MediaMgr_DeviceConfig_Guide.txt /openv/volmgr/bin/acsd /openv/volmgr/bin/acstest /openv/volmgr/bin/avrd /openv/volmgr/bin/driver/64/sg /openv/volmgr/bin/driver/sg /openv/volmgr/bin/goodies/xvmadm /openv/volmgr/bin/ltid /openv/volmgr/bin/mmcrawl /openv/volmgr/bin/odld /openv/volmgr/bin/oprd /openv/volmgr/bin/sg.build /openv/volmgr/bin/sgscan /openv/volmgr/bin/tl4d /openv/volmgr/bin/tl8cd /openv/volmgr/bin/tl8d /openv/volmgr/bin/tl8test /openv/volmgr/bin/tldcd /openv/volmgr/bin/tldd /openv/volmgr/bin/tldtest /openv/volmgr/bin/tlhcd /openv/volmgr/bin/tlhd /openv/volmgr/bin/tlhtest /openv/volmgr/bin/tlmd /openv/volmgr/bin/tpautoconf /openv/volmgr/bin/tpconfig /openv/volmgr/bin/tpformat /openv/volmgr/bin/ts8d /openv/volmgr/bin/tsdd /openv/volmgr/bin/tsdtest /openv/volmgr/bin/vmadm /openv/volmgr/bin/vmcheckxxx /openv/volmgr/bin/vmconf /openv/volmgr/bin/vmd /openv/volmgr/bin/vmdareq /openv/volmgr/bin/vmoprcmd /openv/volmgr/bin/vmquery /openv/volmgr/database/external_densities /openv/volmgr/database/external_drive_mappings /openv/volmgr/database/external_drivetypes /openv/volmgr/database/external_mediatypes /openv/volmgr/database/external_robot_mappings /openv/volmgr/database/external_robotics /openv/volmgr/help/tpconfig/tpconfigdev /openv/volmgr/help/tpconfig/tpconfigrob Problem Description: 4927029 Need NB_34_6 released in Sun format (from 110539-16) 4836227 Need NB_34_5 released in Sun format (from 110539-15) 4759007 need to have NBU 34_4 in Sun format (from 110539-14) 4700664 nbu 3.4 w/patch 110539-13 does not have Solaris 9 listed under add client 4700521 spelling errors in patch 110539-13 README (from 110539-13) 4678476 Single File Restore of FlashBackup 3.4 failed after installing Patch 110539-10 (from 110539-12) 4666899 110539-11 does not have SUNW_PATCH_PROPERTIES in pkginfo (from 110539-11) 4652724 Activity Monitor doens't show jobs on NT client with TRU64Unix Server (from 112065-01) 4499273 September 9th, 2001 problem on jnbSA GUI. (from 112404-02) 4641151 cleaning tape is not recognized as a cleaning tape using LTO driver (from 112404-01) 4641305 add support for HCART/LTO drive types in BusinesServer (from 110539-10) 4523614 110539-09 does not install (from 110539-09) 4432588 stk L180/700 robots not checking in as serialized device w/nbu34,0220 robot code (from 110539-08) 4507040 potential data loss for media servers with multiple nics (from 110539-07) 4503882 110539-06 does not install (from 110539-06) 4500422 Regression issue found for 110539-05 in WOS (from 110539-05) 4324687 Bad management of drives available for backups by NBU Storage Share Option (from 110539-04) 4429632 Bpbackup -i (backup immediate) fails with error 58. (from 110539-03) 4397701 Bpbackup -i (backup immediate) fails with error 25 (from 110539-02) 4387565 Netbackup not working with DES-56 4356569 sg: driver does not support Autosense properly (from 110539-01) 4406098 need veritas patch J0850645 converted to patchadd format Patch Installation Instructions: ------------------------------------------------------------------------- Refer to the Install.info file within the patch for instructions on using the generic 'patchadd' and 'patchrm' scripts provided with each patch. Any other special or non-generic installation instructions should be described below. ------------------------------------------------------------------------- Special Install Instructions: As root on the NetBackup Master Server: (for a NetBackup media server, do steps 1-3) 1) Stop the NetBackup and Media Manager daemons: /usr/openv/netbackup/bin/goodies/bp.kill_all 2) Install patched binaries via patchadd. 3) Restart daemons. /usr/openv/netbackup/bin/initbprd /usr/openv/volmgr/bin/ltid -v If you want to configure and update Solaris 9 clients, execute the following command first: /usr/openv/netbackup/bin/goodies/new_clients You will now be able to add Solaris 9 clients into NetBackup classes. To install software to these clients, follow the instructions in the "Install NetBackup Clients" section of the NetBackup DataCenter Installation Guide - UNIX or NetBackup BusinesServer Getting Started Guide - UNIX. 4) Update the NetBackup clients, including the NetBackup master and media servers, with the update_clients script. /usr/openv/netbackup/bin/update_clients where is one of the following: Solaris Solaris2.6 Solaris Solaris7 Solaris Solaris8 Solaris Solaris9 Solaris Solaris_x86_2.6 Solaris Solaris_x86_7 Solaris Solaris_x86_8 Remember to include the master server's type. Note: This patch contains Solaris client binaries only. If there are other platforms, it is important to patch those clients as well. For the binaries of other clients, download the latest patches from the VERITAS website, http://support.veritas.com. Note: The /usr/openv/netbackup/bin/update_clients command without any parameters will update all the Unix clients. Post-Install Instructions: 1) Make sure that a 3.2 sg driver exists. /usr/sbin/modinfo | grep sg 2) Remove the sg driver. /usr/sbin/rem_drv sg 3) Verify that the sg driver is removed. /usr/sbin/modinfo | grep sg 4) Install new sg driver. /usr/openv/volmgr/bin/driver/sg.install 5) Verify that the sg driver exists and that it is version 3.4. /usr/sbin/modinfo | grep sg 6) postpatch will run an update for NB-Java.tar.Z, Then to update any NetBackup clients that use the NetBackup Java applications. Use /usr/openv/netbackup/bin/update_clients -Install_Java Additional Notes: If nonroot users have been given permission to perform NetBackup or Media Manager administration, (NetBackup System Administrator's Guide, Allowing Nonroot Users to Administer NetBackup page 31 and Media Manager Administrator's Guide, Allowing Nonroot Users to Administer Media Manager, page 9), group and file permissions changes will have to be reapplied as installing this patch will reset the group and file permissions to their delivered value. This can be done by executing: /usr/openv/netbackup/bin/nonroot_admin ============================================ ============= Description Key: ============= ** Description ** Describes a problem that can lead to potential data loss. Please read these problem descriptions carefully. Workaround Any available workarounds to a problem are also listed. Workarounds can be used INSTEAD of applying the patch. Additional Notes Any additional information regarding this problem or feature. ============= Known issues: ============= Description: NetBackup for Lotus Notes does not support the file list directive, ALL_LOTUS_DATABASES. Workaround: Add the Domino data directory, containing the database files with extensions .nsf, .ntf or .box, in the file list instead of using the ALL_LOTUS_DATABASES file list directive. There will be an increase in performance since the NetBackup for Lotus Notes extension does not have to parse the file system for .nsf, .ntf or .box extensions. Description: NetBackup extension for Lotus Notes is not supported in the Microsoft Cluster environment (MSCS). NetBackup does support Lotus Notes in a non-clustered configuration. ============= Current patch ============= Description: ACSD Robotic mounts may fail if many tape drives are loaded at the same time by a single server. Additional Notes: The syslog will contain "waited 60 times for ready" twice for a drive, and then the mount will fail. (All NetBackup Servers) -------------------------------------------------------------------------------- Description: A search for a specific file in an image that is indexed can fail if the file list for the backup contains consecutive files from the same directory. Workaround: Rename the /usr/openv/netbackup/db/images//INDEX directory and do the search again. (All NetBackup Servers) -------------------------------------------------------------------------------- Description: Errors occurred when dismounting media from drives in the ODL library. Error in syslog was: odld[23109]: ODL(1) initiating MOVE_MEDIUM from addr 1 to addr 112, invert = no odld[23109]: ODL(1) Move_medium error, status = 1 odld[23109]: ODL(1) key 5h, ASC 53h, ASCQ 82h Sense key indicates illegal request. Media was used after seeing error with no problems, and drives were used after error with no problem. (All NetBackup Servers) -------------------------------------------------------------------------------- Description: Backups periodically fail with status 81 (Media Manager volume daemon (vmd) is not active). Vmquery and vmcheckxxx/vmupdate may also fail. The bptm logs show one or more entries similar to the following: 19:00:33.946 [5942] <16> string_to_record: cannot convert unrecognized string: 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 NBU Default Rule 19:13:55.448 [6799] <16> string_to_record: expected 63 parameters, only found 10 Workaround: Create the file: /usr/openv/volmgr/VMDB_QUERY_BUFFER_SIZE and set 0 (zero) in the file. (All NetBackup Servers) -------------------------------------------------------------------------------- Description: Tapes might not be unmounted on a media server until all jobs on that media server completed if a user backup was initiated while a non-user backup was active. (All NetBackup Servers) -------------------------------------------------------------------------------- Description: On certain robotic libraries, when robotic auto-cleaning is enabled, drives can be DOWN'd when this process starts. The system log will display the following when this occurs: Mar 31 11:00:47 mtiwbrc05 tldcd[4307]: valid = 1, sel = 10001, barcode = (CLN010 ) Mar 31 11:00:52 mtiwbrc05 tldcd[4307]: TLD(0) I/E offset 0 is not full Mar 31 11:00:52 mtiwbrc05 tldcd[4307]: TLD(0) initiating MOVE_MEDIUM from addr 501 to addr 10 Mar 31 11:00:52 mtiwbrc05 tldcd[4307]: TLD(0) move medium failed because drive is being cleaned, delay and retry No problem, the drive is being cleaned, let's try it again: Mar 31 11:01:53 foo tldcd[4307]: TLD(0) key = 0x5, asc = 0x3b, ascq = 0xe, MEDIUM SOURCE ELEMENT EMPTY Mar 31 11:01:53 foo tldcd[4307]: TLD(0) Move_medium error: CHECK CONDITION Mar 31 11:01:53 foo tldcd[4307]: TLD(0) could not move barcoded tape CLN010 from drive 2 to outport Mar 31 11:01:53 foo tldcd[4307]: TLD(0) closing/unlocking robotic path Mar 31 11:05:23 foo tldcd[4942]: TLD(2) initiating MOVE_MEDIUM from addr 501 to addr 1456 Mar 31 11:05:28 foo tldcd[4942]: TLD(2) closing/unlocking robotic path When recovering from the auto-cleaning, it should not be an error if the drive is now empty. (All NetBackup Servers) -------------------------------------------------------------------------------- Description: Added barcode rule support for API (acs, tlh, tlm, lmf) robots. (All NetBackup Servers) -------------------------------------------------------------------------------- Description: When attempting to read the true image information during a restore, a timeout can occur because an alarm has not been cleared. This will cause the restore to fail. (All NetBackup Servers) -------------------------------------------------------------------------------- Description: If index_clients is run and the clients are already indexed, it just reports "/usr/openv/netbackup/db/INDEXLEVEL exists with a value of 9" and exits. Code was added to reindex the clients if a prompt is responded to in the affirmative. Workaround: Remove the INDEX directory and INDEXLEVEL file and then rerun the command. cd /usr/openv/netbackup/db/ rm INDEXLEVEL rm -rf INDEX (All UNIX NetBackup Servers) -------------------------------------------------------------------------------- Description: When restoring from block incremental images the restore can hang if the first needed backup resides on tape1, the second needed backup resides on tape2, and the third needed backup resides on tape1 again. Workaround: Duplicate the incremental images that reside on the same tape as the full to a different tape. (All NetBackup Servers) -------------------------------------------------------------------------------- Description: A bprd core dump can occur during a restore if FORCE_RESTORE_MEDIA_SERVER or FAILOVER_RESTORE_MEDIA_SERVERS entries exist in the bp.conf file and the failover server name is longer than the original server name. (All NetBackup Servers) -------------------------------------------------------------------------------- Description: If an incoming connection to vmd contained bad or wrong information, the vmd process could die. This fix allows vmd to be more resilient. (All NetBackup Servers) -------------------------------------------------------------------------------- Description: vmd/DA registration processing can crash vmd. The registration string, received by vmd from socket, is copied to a global buffer before passing it to logmsg(). The size of the global buffer (passed to logmsg) is 1024 characters. If the registration string is larger than that, it may overrun the buffer and may cause core dump. (All NetBackup Servers) -------------------------------------------------------------------------------- Description: Increase unload retries for tl8 and ts8 drives. (All NetBackup Servers) -------------------------------------------------------------------------------- Description: NetBackup can now check for TapeAlert conditions from IBM LTO drives. This was accomplished by removing the SCSI mode sense check from the TapeAlert algorithm. The SCSI mode sense check is no longer necessary. (All NetBackup Servers) ================================================================================ ============= 110539-16 ============= Description: If a job fails, is queued again, and then fails because a window closes, the exit status is 196 and indicates the window closed. However, the reason the job failed is because of the previous status. Corrections have been made to ensure the proper status is reported in the job monitor and error logs. Occasionally, errors like this would occur because tapes are supposed to remain mounted after a user backup completes and the media_unmount_delay has passed. Also, if message queues are removed, most of the processes hang making it impossible to kill a job from the job monitor. Processes can now recognize when message queues are removed and terminate gracefully. In addition, the main schedule will signal children processes to terminate immediately. (NetBackup Servers) -------------------------------------------------------------------------------- Description: The NO_TAPEALERT touch file used to disable TapeAlert based drive cleaning was not being used properly in avrd. Partially configured drives (only visible through tpconfig -tables) are no longer scanned by avrd. (NetBackup Servers) -------------------------------------------------------------------------------- Description: An erroneous backup failure and frozen tape could occur if a backup hits EOM, and selects a previously mounted tape to continue the backup. This will only happen if no scratch tapes are available. The error log will contain this message: "FREEZING media id xxxxxx, too many data blocks written, check tape/driver block size configuration" The bptm log will show that a previously mounted tape was selected after EOM processing. (NetBackup Servers) -------------------------------------------------------------------------------- Description: A Customer can successfully install BLI and run an Oracle script-based bli policy without creating the bli scripts. The activity monitor will report a status 0 for the backup. As noted in the bpbkar log below, the process encountered a warning "WRN", however, continued on with the backup and completed with a status 0. Since the block incremental notify script was not found, this backup image would not be a valid image. Without the notify script, the database was not placed into backup mode to obtain a quiesant backup image of the database. bpbkar log 02:00:25 [19616] <2> logparams: bpbkar -r 604800 -ru root -dt 93610 - to 0 -clnt as370 -class as370_BLI -sched incrementals -st INCR - bpstart_to 300 -bpend_to 300 -read_to 300 -block_incr -keyword spps -streams 1 -bt 990511247 - prev_bli_incr 990417647 -last_bli_full 990331247 -use_otm -fso -b as370_0990511247 -kl 14 -shm 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (ufs) /dev/vx/dsk/rootvol on / 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (ufs) /dev/vx/dsk/usr on /usr 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (proc) /proc on /proc 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (fd) fd on /dev/fd 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (ufs) /dev/vx/dsk/var on /var 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (vxfs) /dev/vx/dsk/appdg/app_01 on /app1 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (vxfs) /dev/vx/dsk/appdg/app_02 on /opt/openv 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (autofs) - hosts on /net 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (autofs) auto_home on /home 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (autofs) - xfn on /xfn 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (NFS) as370:vold(pid620) on /vol 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (vxfs) /dev/vx/dsk/appdg/app_03 on /app1/inktomi 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (vxfs) /dev/vx/dsk/Foracle_dg1/spps1_vol on /oracle_spps1 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (vxfs) /dev/vx/dsk/Foracle_dg2/arch_vol on /oracle_spps/arch 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (vxfs) /dev/vx/dsk/Foracle_dg1/spps3_vol on /oracle_spps3 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (vxfs) /dev/vx/dsk/Foracle_dg1/spps2_vol on /oracle_spps2 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (vxfs) /dev/vx/dsk/Foracle_dg1/oracle_vol on /opt/oracle 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (vxfs) /dev/vx/dsk/Foracle_dg1/data_vol on /oracle_spps/data 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (vxfs) /dev/vx/dsk/Foracle_dg3/dbf_vol on /opt/oracle/admin/spps/dbf 02:00:25 [19616] <2> bpbkar initialize: INF - Selecting files modified or changed after Mon May 21 00:00:15 2001 02:00:25 [19616] <4> bpbkar main: INF - setenv KEYWORD=spps 02:00:25 [19616] <4> bpbkar main: INF - setenv STREAM_PID=19616 02:00:25 [19616] <4> bpbkar main: INF - setenv STREAM_NUMBER=0 02:00:25 [19616] <4> bpbkar main: INF - setenv STREAM_COUNT=0 02:00:25 [19616] <4> bpbkar main: INF - setenv STREAMS=1 02:00:25 [19616] <4> bpbkar main: INF - setenv BPSTART_TIMEOUT=300 02:00:25 [19616] <4> bpbkar main: INF - setenv BPEND_TIMEOUT=300 02:00:25 [19616] <4> bpbkar main: INF - setenv BACKUPID=as370_0990511247 02:00:25 [19616] <4> bpbkar main: INF - setenv UNIXBACKUPTIME=0990511247 02:00:25 [19616] <4> bpbkar main: INF - setenv BACKUPTIME=Tue May 22 02:00:47 2001 02:00:25 [19616] <4> bpbkar main: INF - Inform when done 02:00:25 [19616] <8> bpbkar notify: WRN - Missing block incremental notify script /usr/openv/netbackup/bin/bpstart_notify.as370_BLI 02:00:25 [19616] <4> bpbkar main: INF - Could not open /usr/openv/netbackup/bin/BLOCK_INCR/class_in_control.spps. Setting CLASS_IN_CONTROL to as370_BLI 02:00:25 [19616] <4> bpbkar main: INF - BACKUP START 02:00:25 [19616] <2> bpbkar main: INF - CONTINUE BACKUP message received 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (ufs) /dev/vx/dsk/rootvol on / 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (ufs) /dev/vx/dsk/usr on /usr 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (proc) /proc on /proc 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (fd) fd on /dev/fd 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (ufs) /dev/vx/dsk/var on /var 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (vxfs) /dev/vx/dsk/appdg/app_01 on /app1 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (vxfs) /dev/vx/dsk/appdg/app_02 on /opt/openv 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (autofs) - hosts on /net 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (autofs) auto_home on /home 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (autofs) - xfn on /xfn 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (NFS) as370:vold(pid620) on /vol 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (vxfs) /dev/vx/dsk/appdg/app_03 on /app1/inktomi 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (vxfs) /dev/vx/dsk/Foracle_dg1/spps1_vol on /oracle_spps1 02:00:25 [19616] <2> mount build_mount_list: INF - Processing (vxfs) /dev/vx/dsk/Foracle_dg2/arch_vol on /oracle_spps/arch Workaround: Create the bli scripts using the setup_bli_scripts utility found in the /usr/openv/netbackup/ext/db_ext/oracle/bin directory. (NetBackup Clients: Solaris2.6) -------------------------------------------------------------------------------- Description: When NetBackup DB indexing is turned on, some searches will return "EXIT STATUS 227: no entity was found" when the same path exists sequentially in the index file with the .i1 suffix. Workaround: Rename the /usr/openv/netbackup/db/images//INDEX directory so that the index files cannot be used. (NetBackup Servers) -------------------------------------------------------------------------------- Description: Hard links with path names greater than 100 characters, in which the linked file also has a path name greater than 100 characters, could not be restored individually without also restoring the linked file. Workaround: Restore both the hard link file and it's linked file. (NetBackup Unix Clients) -------------------------------------------------------------------------------- Description: When searching an image that is indexed, the search can fail to return the contents of a directory. This failure occurs if the following is true: 1. The directory is a file system mountpoint and the mountpoint is part of another file system. 2. The parent file system was backed up before the file system that is being searched for. 3. The last entry in the parent file system was the directory mountpoint. Workaround: Turn off indexing or add "/*" to the end of the search directory. (NetBackup Servers) -------------------------------------------------------------------------------- Description: Cleaning media can be overused when cleaning does not complete within 30 minutes after a cleaning request is initiated. After 30 minutes, the cleaning request is removed from internal tables, so that when the cleaning actually does complete, the cleaning statistics (last clean time, total mount time, number of cleanings remaining for the cleaning tape) are not updated. Also, when the cleaning request is removed, system log messages such as "Removed cleaning queue entry for drive index 1 (drive 2), it has not completed in 30 minutes" show up, giving a false impression that the drive was not cleaned, when in fact the cleaning request may have been still active. The following vm.conf entry is available to make the cleaning request timeout configurable: CLEAN_REQUEST_TIMEOUT = minutes where "minutes" specifies how long Media Manager should wait for a drive to be cleaned before removing the request from the cleaning queue if the request has not been processed. The minimum for minutes is 1. The maximum is 144000 (100 days). The default is 30. 0 is converted to the default value of 30. (NetBackup Servers) -------------------------------------------------------------------------------- Description: Suppose you have a machine that is both a NetBackup server and client, and that same machine has more than 20 network interfaces with a NetBackup client defined for each interface. If the machine (as a server) tries to connect to itself via one of the network interface clients, the connection may fail with a "server not allowed access" error (status 46) or "access to the client was not allowed" error (status 59). For example, the machine may be a media server that is trying to backup a client that is defined for one of its network interfaces. It is possible that you will receive a status 59 error. Workaround: Add the hostnames of all the network interfaces to the SERVER list in the machine's configuration. (NetBackup Servers and Clients) -------------------------------------------------------------------------------- Description: When using multiple data streams with NDMP, the dumplevels for incrementals were sometimes incorrect. The result was more data being backed up that should be. (NetBackup Servers) -------------------------------------------------------------------------------- Description: When doing TIR backups with move detection, files that have not changed may be backed up during an incremental. In extreme cases, whole directories may be backed up even when none of the files have changed. (NetBackup Unix Clients) -------------------------------------------------------------------------------- Description: Configuration of TIMEOUT_IN_QUE did not work. Jobs would remain queued even after the timeout had expired. (NetBackup Servers) -------------------------------------------------------------------------------- Description: When specifying a preprocess interval using bpconfig or bpadm (Unix only), there is not a way to set the default value. In addition, the System Administrator's Guide has conflicting information on what a value of 0 indicates. In one place in the document it states that 0 indicates the preprocess interval should occur prior to every backup. It another place it state that a value of 0 should reset the parameter to the default value (every 4 hours). Use the following command to set the default preprocess interval (every 4 hours): bpconfig -prep -1 Use the following command to set the preprocess interval to occur prior to every backup: bpconfig -prep 0 In addition, bpadm is changed to allow these options to be set, and both bpadm and bpconfig clearly report the settings. Workaround: To reset the default preprocess interval, use bpconfig to set the value to 4 hours as follows: bpconfig -prep 4 To set the preprocess interval to occur prior to every backup, use a preprocess value of 0: bpconfig -prep 0 (NetBackup Servers) -------------------------------------------------------------------------------- Description: NCVU 4.5.3.1 Combined the three NCVU binaries (NCVU, NCVU_media_server and NCVU_client) in one binary (NCVU). This reduces the footprint and download times by 66%. This required adding a new command line parameter -host_node . This also negated the need for the chk_nbu_node.pm module. (NetBackup Servers) -------------------------------------------------------------------------------- Description: During a restore of directories with both ACL (Access Control List) entries and the SETGID bit set, the SETGID bit would not get set. (NetBackup Unix Clients) -------------------------------------------------------------------------------- Description: The bpdbm -consistency [1 | 2] command may incorrectly report errors for database extension backups. The errors will be reported as follows: checking image file >>PRIMARY_COPY is set to an invalid copy >>EXPIRATION is not set to the next valid copy to expire Workaround: Ignore these errors for database extension backups. (NetBackup Servers) -------------------------------------------------------------------------------- Description: A customer, who has a large number of tapes in their volume database (80,000), encountered a situation where they attempted to change a pool of 287 tapes in one operation. This request took vmd 18 minutes to complete, and caused backup jobs to timed out while waiting for vmd to respond. Workaround: Select smaller numbers of volumes to perform the change. For example, instead of changing a pool of 200 tapes in one operation, perform the operation on 4 groups of 50 volumes. (NetBackup Servers) -------------------------------------------------------------------------------- Description: Problems with failing backups, restores, and tape unmounting have been observed on AIX systems using 6227/6228 Fibre Channel HBA's. The problems can occur when your tape devices are connected via a Fibre Channel Bridge. The AIX system is sending Fibre Channel PLOGI frames to tape devices when the tape path is opened by an application. This PLOGI frame can cause the application to see an I/O error on a subsequent command. This failing command can sometimes cause the above-mentioned problems. If you suspect that this is causing problems on your server, you may use the file "/usr/openv/volmgr/database/AIX_OPEN_HBA" and restart the media manager daemons. Look for messages such as, "avrd[30220]: Drive /dev/rmt57.1, parent /dev/fscsi0 opened" in the syslog to verify that the patch is working correctly. Every 6227/6228 HBA in your system that has tape devices attached to it, should have a message in the syslog. You may be able to monitor your bridge and/or fabric equipment to detect the PLOGI frames. (NetBackup Servers: RS6000) -------------------------------------------------------------------------------- Description: If a Unix machine is running osinetd, our attempt to signal inetd with a HUP will accidentally kill osinetd. (NetBackup Unix Clients) -------------------------------------------------------------------------------- Description: Restoring using RMAN and getting error 5. The files are retrieved and written to disk and then failure occurs. (All UNIX NetBackup Servers) -------------------------------------------------------------------------------- Description: Performing an import or verify of an image that is created from the backup of a Windows client on a UNIX media server could cause a core dump in tar. Workaround: Perform the import or verify of Windows images on a Windows media server. Additional Notes: Added new potential warning message: "Skipping unknown extras header %s." (NetBackup Unix Clients) -------------------------------------------------------------------------------- Description: Add device discovery support for the Quantum DX30 robot. (NetBackup Servers) -------------------------------------------------------------------------------- Description: Searching for backup images to duplicate (eg., bpduplicate -PM) may periodically fail with a 43 (unexpected message received) or 144 (invalid command usage). (NetBackup Servers) -------------------------------------------------------------------------------- Description: The first Lotus transaction log extent in a backup image has incorrect owner and group permissions. Rather than having the owner and group permissions for the Lotus Domino server administrator, the owner and group are root and root. (NetBackup Unix Clients) -------------------------------------------------------------------------------- Description: When the tpcommand debug directory exists, calls to tpconfig -robdrtype and a couple of others may dump core. This includes hidden calls to said commands such as vmupdate uses. (NetBackup Servers) -------------------------------------------------------------------------------- Description: When a tape fills and spans to another tape, and a new tape is used, scheduler was unaware that the original drive became available for reuse. The would cause scheduler to not use all the drives available. (NetBackup Servers) -------------------------------------------------------------------------------- Description: Import of compressed files of size (2GB - 1MB) through 2GB won't restore. Workaround: Manually change the size of the file in the image database to the correct value (reduce size by 1GB). (NetBackup Unix Clients) -------------------------------------------------------------------------------- Description: When searching a backup that is indexed and the search path is at a lower level that what was actually backed up, the search can get into an infinite loop. (NetBackup Servers) -------------------------------------------------------------------------------- Description: If a SSO drive Release message fails, the server can continue to scan drive as it is used by other systems. This can cause the media to be rewound during a backup or restore operation. (NetBackup Servers) ============= 110539-15 ============= Description: A potential problem existed in the socket connection between bptm and bprd. If bptm took a long time to process the XFERBLOCKS that were sent to bptm from bprd, bptm could get a 'connection reset' error during restore. bprd will close the socket immediately after sending the XFERBLOCKS. We should safely shutdown the socket without fully closing it so that bptm will not receive a 'connection reset' error if it attempts to communicate on that socket. Additional Notes: This only effects file restores on FlashBackups, which use single file restore (SFR) capabilities. It seems this is more likely to occur if restoring large, sparse files from a FlashBackup image. Also, a tape that is heavily multiplexed may be the source of this problem and result in many more XFERBLOCKS in bptm, causing it to delay the restore process. -------------------------------------------------------------------------------- Description: If a TapeAlert-based cleaning was initiated by ltid or avrd for an ACS, TLH, or an LMF drive, the drives would never get released by the vmd/DA. Since these robot types clean themselves, we do not need to initiate cleaning. If TapeAlert-based cleaning was initiated by ltid or avrd for non-API SSO robots and no cleaning media was available, or the number of cleanings was 0, the drives would never get released by the vmd/DA. Additional Notes: Added a touchfile for disabling TapeAlert based cleaning. For Unix, this is /usr/openv/volmgr/database/NO_TAPEALERT. For Windows, this is volmgr\database\NO_TAPEALERT. -------------------------------------------------------------------------------- Description: sgscan and sg.build would only detect tape and library devices connected to Sun F/C HBA's. If a bridge is exporting a data mover, or other device on LUN 0, then these utilities wouldn't configure the bridge for use with the VERITAS SG driver. Workaround: Manually add the bridge to the sg.conf and devlink.tab files. -------------------------------------------------------------------------------- Description: After a media write error (error 84) from bptm during a multiplexed duplication, the bpduplicate process will hang. But the bptm daemons exit properly. Here is a snapshot of admin log after this error. 22:07:44 [167.55] <2> read_media_msg: ERROR 84 22:07:44 [167.55] <2> process_WRITE_media_msg: ? 22:07:44 [167.55] <4> process_WRITE_media_msg: got ERROR 84 from media manager 22:07:44 [167.55] <2> terminate_mpx_group: called by write 22:07:44 [167.55] <2> send_WRITE_media_msg: (pid 279) TERMINATE 22:07:44 [167.55] <16> put_long: (11) network write() error: Connection reset by peer.; socket = 88 22:07:44 [167.55] <16> send_WRITE_media_msg: could not write TERMINATE to media socket. bptm pid = 279 22:07:44 [167.55] <2> send_READ_media_msg: (pid 173) STOP ALL 22:07:44 [167.55] <2> signal_media: setting pending msg event 22:07:44 [167.55] <2> release_write: ? 22:07:44 [167.55] <2> handle_backup_image_status: hcus-ny24_1016154366 status = 84 22:07:44 [167.55] <2> getsockconnected: host=HCUS-NY24 service=bpdbm address=10.40.33.12 protocol=tcp non-reserved port=13721 22:07:44 [167.55] <2> getsockconnected: host=hcus-ny24 service=bpcd address=10.40.33.12 protocol=tcp reserved port=13782 22:07:44 [167.55] <2> getsockconnected: Connect to hcus-ny24 on port 526 22:07:44 [167.55] <16> BPDUPLICATE: host hcus-ny24 backupid hcus- ny24_1016154366 write failed, media write error (84). 22:07:44 [167.55] <2> duplication_complete: hcus-ny22_1016154357 finished:1 write_finished:1 22:07:44 [167.55] <2> duplication_complete: hcus-ny22_1016154358 finished:1 write_finished:1 22:07:44 [167.55] <2> duplication_complete: hcus-ny22_1016154359 finished:1 write_finished:1 22:07:44 [167.55] <2> duplication_complete: hcus-ny23_1016154360 finished:1 write_finished:1 22:07:44 [167.55] <2> duplication_complete: hcus-ny23_1016154361 finished:1 write_finished:1 22:07:44 [167.55] <2> duplication_complete: hcus-ny23_1016154362 finished:1 write_finished:0 22:07:44 [167.55] <2> handle_backup_image_status: duplication not complete After this point, bpduplicate will loop for statuses indefinitely. 03:09:40 [167.55] <2> mon_mm: select: 0, w:0, r:-1, h:0 03:09:40 [167.55] <2> start_indep_grp: active_mpx_groups:0, active_backups:0, Starting 03:09:40 [167.55] <2> start_indep_grp: p_image = NULL 03:09:40 [167.55] <2> start_indep_grp: oldest_time = 0x7fffffff 03:09:40 [167.55] <2> start_indep_grp: active_mpx_groups = 0 03:09:40 [167.55] <2> start_indep_grp: p_active_dup_for_first_grp = NULL 03:09:40 [167.55] <2> start_indep_grp: p_first_mpx_group_for_restore = NULL 03:09:40 [167.55] <2> next_active_mpx_group: 8573296: bptm_pid:0 03:09:40 [167.55] <2> next_active_mpx_group: 8573216: bptm_pid:0 03:09:40 [167.55] <2> next_active_mpx_group: 8573376: bptm_pid:0 03:09:40 [167.55] <2> next_active_mpx_group: 8573296: bptm_pid:0 03:09:40 [167.55] <2> next_active_mpx_group: 8573216: bptm_pid:0 03:09:40 [167.55] <2> mon_mm: called by mpxdup 03:09:50 [167.55] <2> mon_mm: select: 0, w:0, r:-1, h:0 03:09:50 [167.55] <2> start_indep_grp: active_mpx_groups:0, active_backups:0, Starting 03:09:50 [167.55] <2> start_indep_grp: p_image = NULL 03:09:50 [167.55] <2> start_indep_grp: oldest_time = 0x7fffffff 03:09:50 [167.55] <2> start_indep_grp: active_mpx_groups = 0 03:09:50 [167.55] <2> start_indep_grp: p_active_dup_for_first_grp = NULL 03:09:50 [167.55] <2> start_indep_grp: p_first_mpx_group_for_restore = NULL 03:09:50 [167.55] <2> next_active_mpx_group: 8573296: bptm_pid:0 03:09:50 [167.55] <2> next_active_mpx_group: 8573216: bptm_pid:0 03:09:50 [167.55] <2> next_active_mpx_group: 8573376: bptm_pid:0 03:09:50 [167.55] <2> next_active_mpx_group: 8573296: bptm_pid:0 03:09:50 [167.55] <2> next_active_mpx_group: 8573216: bptm_pid:0 03:09:50 [167.55] <2> mon_mm: called by mpxdup ..... -------------------------------------------------------------------------------- Description: Device discovery support has been added for the following devices: NEC~LTO~Autoloader NEC LL0101H-0A NEC~SuperDLT~Autoloader NEC SL0101H-0S SONY~LIB-81 SONY LIB-81 Workaround: Manually configure these libraries. The NEC libraries are both TLD and the Sony library is a TL8. -------------------------------------------------------------------------------- Description: AVRD would scan a drive as it was becoming ready, which could cause a tape change to go un-noticed. AVRD was not checking if the tape was write protected. All tapes where reported as writeable. AVRD could not send the startup SCSI Release, because the tape path could not be opened. -------------------------------------------------------------------------------- Description: Restores of Quick I/O files to a file system with a larger block size are allocated more blocks than necessary. The extent information needs to be modified to account for the larger file system block size. Workaround: Restore to a file system with the same block size. -------------------------------------------------------------------------------- Description: Incorrect "Bad image header" entries occur in the problem report and bpdbm debug log when block-level, incremental images expire. The bpdbm log shows entries similar to the following: 10:49:02.347 [6060] <2> db_get_image_info: fopen (/usr/openv/netbackup/db/images/hartstongue/1018000000/test_1018363428_FULL) : No such file or directory (2) 10:49:02.348 [6060] <2> image_file_name: ? 10:49:02.349 [6060] <16> delete_expired_backups: Bad image header: test_1018363428_FULL -------------------------------------------------------------------------------- Description: Add SONY DMS libraries to external mappings file, as TLD robots. Ignore storage slots that are configured as cleaning slots for automatic robotic cleaning. -------------------------------------------------------------------------------- Description: Certain robots classified as TL8, which return read element status data greater than 52 bytes for the drive element descriptors, do not work on UNIX. There are two external symptoms to this problem: 1. In tl8test, when an 's d' is entered to perform a read element status on the drive elements, the addr field is incorrect: /usr/openv/volmgr/bin/tl8test -r /dev/sg/c2t0l0 Opening /dev/sg/c2t0l0 Enter tl8 commands (? returns help information) s d drive 1 (addr 224) access = 1 Contains Cartridge = no SCSI ID from drive 1 is 1 drive 2 (addr -1234) access = 1 Contains Cartridge = no SCSI ID from drive 2 is 2 READ_ELEMENT_STATUS complete 2. The system event log will state that the drive does not exist in the robot. Additional Notes: This has been observed on the CPQ SSL2020 robot. -------------------------------------------------------------------------------- Description: NetBackup catalog backups error with a "media position error" (error code 86) on a Tru64 server. The admin log shows an error similar to the following: 11:29:31 [28134] <2> io_ioctl: command (2)MTBSF 1 on file /tmp/sync_DBA113 11:29:32 [28134] <2> child_wait: SIGCLD: exit=0, signo=0 core=no ... 11:29:32 [28134] <16> io_ioctl: ioctl (MTBSF) failed on file DBA113, Interrupted system call -------------------------------------------------------------------------------- Description: Resume logic must re-establish the SCSI Reservation prior to checking the current tape position. The problem that caused resume logic to be activated may have caused the SCSI Reservation to be lost. -------------------------------------------------------------------------------- Description: When running AUTO_UPDATE_ROBOT, the tldcd parent process can core dump while attempting to empty out the media access port on some robots. Workaround: Turn off AUTO_UPDATE_ROBOT -------------------------------------------------------------------------------- Description: TIR with move detection fails to manage the tir_info directory properly when wildcards are used in the filelist. This causes the incrementals to back up all the files specified each time an incremental is run for the filelist entries with a wildcard in the name. Workaround: Don't use wildcards in the filelist when using TIR with move detection. -------------------------------------------------------------------------------- Description: Firmware changes to the Breece Hill Q140 caused the inquiry string to change. By changing the inquiry string, the device in the robotic control daemon was no longer recognized. Workaround: Download the latest mappings files from the NetBackup support page. -------------------------------------------------------------------------------- Description: If there are two libraries (example: A and B) configured on the same server. This would cause B to try and behave as A and possibly fail, if B was not capable of doing what A does. For example, if A is capable of doing an "extend and retract" and B is not. Then if B tries to do an "extend and retract" it would fail. This is not the case with all libraries, just a few of them. -------------------------------------------------------------------------------- Description: Incorrect protocol detection causes tldcd and/or tldtest to hang. The syslog may contain an error like "Control daemon connect or protocol error" when this occurs. Device configuration will work. TLD libraries will not function and inventory or tape movement will be impossible. Workaround: Touch file /usr/openv/volmgr/TLDx_SCSI, where x is the robot number, or 99 for tldtest. -------------------------------------------------------------------------------- Description: Media is unmounted after a restore prior to the media unmount delay interval. For example, the bptm log has entries similar to the following: 10:11:04 [1935] <4> mpx_read_backup: successfully restored 1 of 1 requests, read total of 107892 Kbytes at 7269.448 Kbytes/sec 10:11:04 [1935] <2> read_backup_unmount_delay: waiting 180 seconds before unmounting media after restore 10:11:05 [135] <2> bptm: INITIATING: -U 10:11:05 [135] <2> db_byid: search for media id AA1010 10:11:05 [135] <2> db_byid: AA1010 found at offset 6 10:11:05 [135] <2> tpunmount_all: tpunmount'ing /usr/openv/netbackup/db/media/tpreq/AA1010 10:11:05 [135] <2> bptm: EXITING with status 0 <---------- -------------------------------------------------------------------------------- Description: There are long device file names for devices in fabric fiber channel. Patch NB_34_1 introduced support of path names that looked like the following: "/dev/rmt/1234567890123456/lun1nrv/c2p1" However, medium length path names similar to the following could not be handled: "/dev/rmt/1234567890123456/lun1nrv/c2p1fffffffff" This patch adds support to handle medium length path names. NetBackup 4.5 has the ability to handle the longest form of path name, that takes the following form: "/dev/rmt/1234567890123456/lun1nrv/c2p1234567890123456". Workaround: Connect devices in a loop instead of fabric. -------------------------------------------------------------------------------- Description: If running bpcd with the -standalone option (rather than running bpcd through inetd or xinetd) on a Linux client, server connections to bpcd on the client may be delayed or timed out. Typically, this happens during a client initiated restore. A "cannot connect on socket" error on the server is likely. Workaround: Use the default method of running bpcd through inetd or xinetd. -------------------------------------------------------------------------------- Description: During NDMP restores, process bpbrm would hang indefinitely if bp.conf contains MEDIA_UNMOUNT_DELAY = 0. -------------------------------------------------------------------------------- Description: Added device discovery support for the IBM 3607 autoloader. Workaround: manually configure this device as a TLD -------------------------------------------------------------------------------- Description: On UNIX clients with a local file system mounted as a subdirectory of a loopback file system (lofs), that local file system cannot be backed up by a normal NetBackup file system backup. For example, assume the following mounted filesystems: /local/2 (local mount of /dev/dsk/part2) /opt/document (lofs, looped back to /local/2/opt/document) /opt/document/test (local mount of /dev/dsk/partt) When attempting to backup /opt/document/test, NetBackup will report a successful backup. However, it will not backup the data under /opt/document/test. Workaround: Either unmount the loopback filesystems before backing up the local filesystem, or use raw partition (or FlashBackup) to backup the disk partition. -------------------------------------------------------------------------------- Description: On NDMP backups, an asterix in a set command would trigger an attempt to autodiscover on the client. This would cause the backups to fail with a 198 error code. Workaround: If NEW_STREAMS are used to specify the streams, autodiscover will not attempt to expand the wild card. -------------------------------------------------------------------------------- Description: On WinNT, there is a 2K command line length limitation. If the GUI or CLI generates a bpclinclude command line that exceeds that length, it fails. Additional Notes: This fix adds "bpclinclude -f " functionality. -------------------------------------------------------------------------------- Description: Add device support for the COMPAQ TSL-10000 DAT Autoloader. -------------------------------------------------------------------------------- Description: Changes to sort the files in a directory can cause non-optimal disk access patterns when there are large numbers of small files. Performance during a backup is much better if the files are backed up in the order they were read from the directory. -------------------------------------------------------------------------------- Description: Increased the maximum number of drives that may be configured in a TLH robot from 64 to 127. -------------------------------------------------------------------------------- Description: Add support for new ACSLS 6.1 media and drive types. -------------------------------------------------------------------------------- Description: Possible vmd core dump when using the command line to change the number of days to keep media manager debug logs. For example: vmquery -keep_logs 1000 Workaround: Manually edit the "DAYS_TO_KEEP_LOGS = " entry in the vm.conf file. -------------------------------------------------------------------------------- Description: Restore using RMAN receives error 5. The failure occurs after the datafiles files are retrieved and written to disk. -------------------------------------------------------------------------------- Description: Allow NetBackup 3.4 server to act as a GDM-managed server with a NetBackup 4.5 master server. Additional Notes: For Solaris, HP 11.0, DEC Alpha, and IBM servers, if you intend to manage your 3.4 server with GDM 4.5, you will need to modify system startup scr ipts so that they start the nbdbd and visd daemons when the system boots and terminate them at system shutdown. Always start the Media Manager daemons before the NetBackup daemons. In your initialization scripts for media servers, start and stop only ltid. Do not start and stop bprd from media-server initialization scripts. All server platforms have various scripts in the following directory: /usr/openv/netbackup/bin/goodies For DEC Alpha and Solaris 2.x, the goodies directory contains the S77netbackup and K77netbackup scripts. S77netbackup starts the NetBackup and Med ia Manager daemons and K77netbackup stops the NetBackup and Media Manager daemons. For Solaris 2.x place these scripts in the server's /etc/rc2.d (startup) and /etc/rc0.d (shutdown) directories. For DEC Alpha, place them in the /sbin/rc3.d (startup) and /sbin/rc0.d (shutdown) directories. For HP 11.0, the goodies directory contains the S777netbackup and K77netbackup scripts. S777netbackup starts the NetBackup and Media Manager daemo ns and K77netbackup stops the NetBackup and Media Manager daemons. Place these scripts in the server's /etc/rc2.d (startup) and /etc/rc0.d (shutdo wn) directories. For IBM servers, the goodies directory contains the rc.veritas.aix script. This script starts the Media Manager and NetBackup daemons. Place this script into the server's /etc directory and call it during the level 2 boot process. Edit the server's /etc/inittab file and add this line: veritas:2:wait:/etc/rc.veritas.aix Call this script after most of the other level 2 inittab entries like rctcpip and diagd. -------------------------------------------------------------------------------- Description: During certain incremental backups, the file list would not always include all the directories required to find the correct file. This could cause bplist to fail (with a status 227 -- "no entity was found"), or certain files to not appear in the restore list. Workaround: The bplist error could be worked around by not appending a "/" at the end of the path (for a directory), or by turning off indexing. -------------------------------------------------------------------------------- Description: Background unmounts from the scheduler were hanging because the scheduler was waiting on process ID's under NT. Process ID's could be reused and the scheduler was waiting on the wrong process to terminate. The scheduler can now monitor background unmounts using process handles. -------------------------------------------------------------------------------- Description: Backup verification fails with "unexpected message received" error. A restore of the backup is successful. The admin log will have an entry similar to the following: 09:56:51 [19428] <16> get_adaptable_string: (4) network read() error: Connection reset by peer (131) -------------------------------------------------------------------------------- Description: Enhanced tape drive selection for NDMP backups and restores when multiple NAS appliances are sharing the same robot. -------------------------------------------------------------------------------- Description: Trouble can occur when you try to restore files on Mac OS X if the files were backed up with the compression option (not the default) selected. The first file that includes an HFS resource fork will cause error messages similar to the following to occur in the progress log: 13:26:57 (4.001) /Volumes/Partition 5/backup/Library/Application Support/DVD Player Scripts/03Go To Chapter....scpt was not restored. After the first "was not restored" file, all subsequent files will not be restored. -------------------------------------------------------------------------------- Description: If multiple network interface cards are used on a master server, MEDIA_UNMOUNT_DELAY would not work. This was because the names of the host would not compare correctly. The test has been changed to use the primary host name for the compare so unmounts would not be sent to a media server that had a pending delayed unmount. bpsched would stop processing for extended periods of time if a media server is powered down or disconnected from the network. The timeout for a connect is quite long so signals are now used to timeout the connect based on slave_connect_timeout. Also, the ping used in the query to bptm for the number of available drives would attempt to contact the media server even after it was determined to be down which caused multiple timeouts to occur. This attempt is now done only once. -------------------------------------------------------------------------------- Description: Improve drive selection for NDMP restores by allowing robotic daemons to select any available drive in the tape library and not be required to select a drive connected to a specific NDMP host. Also correct a bpsched problem where backups were being started on a storage unit with no available drives, resulting in error code 20 (Invalid argument). -------------------------------------------------------------------------------- Description: Enhanced the robotic drive selection algorithm for the Shared Storage Option (SSO) to help avoid tape mount delays. The robotic daemons previously checked with ltid, the Media Manager device daemon and NetBackup Device Manager service. Now, the robotic daemons make the check directly for assigned drives, skipping those, thereby reducing the load on ltid. -------------------------------------------------------------------------------- Description: Resources busy errors, (134 errors), should not show up in the error log and should not generate mail. There were also some socket errors logged because bpbrm did not recognize that the scheduler child process had terminated because of the 134 error. -------------------------------------------------------------------------------- Description: The SAP backup fails when running SAP 3.4_4 with 4.5 server. -------------------------------------------------------------------------------- Description: Connections to vmd intermittently fail if vmd does not receive notification of the connection within a timeout period. Messages like the following may appear in the bptm log when this happens: 23:30:57 [24351] <16> get_long: (1) cannot read (byte 1) from network: Interrupted system call (4) 23:30:57 [24351] <16> getrequestack: unable to get request ack string: Interrupted system call (4), stat=-2 23:30:57 [24351] <16> vmdb_query_byID_getpool: unable to send request to server: network protocol error (39) -------------------------------------------------------------------------------- Description: vmd logs too much information to the daemon debug log by default. Every string that is sent and received over the socket is logged. In cases where there are tens of thousands of volumes in the volume database, the daemon debug log can hit the current 2GB size limit. Much of the information logged is unneeded or only needed when trying to debug a rare problem. By reducing the volume of logging it will be easier to diagnose problems. -------------------------------------------------------------------------------- Description: When the volume database contains tens of thousands of volume records, some vmd queries did not return the list of volume records to the caller in the most efficient manner possible. Whenever one of these queries was made, it would consume the vmd connection for an extended period of time, leading to contention for vmd connections among multiple processes and machines in large NetBackup installations. This change replaces many of those queries with a more efficient version which will both speed up the return of the list of volume records to the caller and, in turn, reduce the contention for vmd connections in large NetBackup installations. ================================================================================ ============= 110539-14 ============= === Note: now supporting Solaris 9 as Server and Client === ================================================================================ Description: Restores of machines that cannot be accessed from the master server fail with status code 56. To get the behavior where the client is first checked to make sure it is reachable, create a file named CHECK_RESTORE_CLIENT in the netbackup directory. This will signal that the client should be checked before a restore is attempted to prevent an unresponsive client from slowing down other restores (waiting for timeouts to occur) when the other client's data resides on the same tapes. To create the CHECK_RESTORE_CLIENT file on UNIX, just do the following command as root on the master server: touch /usr/openv/netbackup/CHECK_RESTORE_CLIENT If you do not want the restore client to be checked, make sure this file does not exist on the master server. Workaround: Add the client's IP address on the master so that it is reachable. (NetBackup Servers) -------------------------------------------------------------------------------- Description: The problem is that "MEDIA_ID_BARCODE_CHARS" would only work if the vm.conf had an entry for every different barcode length in the robot. This fix gets rid of the problem. The user could now add only one rule if they wish, and that rule will work for that barcode length. And the media id's for the rest of the media would get generated as before. Workaround: If the user adds n "MEDIA_ID_BARCODE_CHARS" rules for n different barcode lengths of media in the robot. Then this feature works correctly. (NetBackup Servers) -------------------------------------------------------------------------------- Description: An active restore job will have the status changed to 50. This occurs at midnight it the job has been active for more that 48 hours. Additional Notes: The status is changed to 50, but the job is still active. When the job completes the status will be set to 0. (NetBackup Servers) -------------------------------------------------------------------------------- Description: Qualstar model 88264 cannot use mailslot for media eject. Also, most Qualstar libraries will display a negative mailslot address in tl8test and tldtest. (NetBackup Servers) -------------------------------------------------------------------------------- Description: When trying to run update_clients to push or install patch binaries to MacOSX or MacOSXS clients, update_clients fails saying it cannot update non-UNIX clients. (NetBackup Servers) -------------------------------------------------------------------------------- Description: A rare end-case in vmd socket processing can cause file descriptors to be consumed and not released when handling network requests. The eventual side-effect is for vmd to log errors when opening files with the message "too many open files". This situation takes place after network activity has caused vmd to consume the maximum number of file descriptors permitted by the operating system. Workaround: Stop and restart vmd. (NetBackup Servers) -------------------------------------------------------------------------------- Description: A race condition may occur in ltid, the Media Manager device daemon and NetBackup Device Manager service, when a remote drive scan, drive assignment, or drive control command unexpectedly fails. This applies only to configurations with the Shared Storage Option. During the race condition, ltid continually retries the remote operation, terminates the existing rdevmi (remote device management interface) process, creates a new rdevmi process on the remote host, and re-registers shared drives with vmd/DA, the device daemon and NetBackup Volume Manager service. If debug logs are enabled, the volmgr "daemon" debug may show continued attempts to start new rdevmi process: 10/17/01 16:42:04 <0> process_requests[26516]: oprd received string 56 -1 0 0 DECTZ89CDEC0 *NULL* *NULL* 10/17/01 16:42:05 <0> process_requests[26516]: received TERMINATE request 10/17/01 16:42:05 <1> rdevmi[26719]: INITIATING 10/17/01 16:42:05 <0> rdevmi[26719]: got CONTINUE, connecting to ltid 10/17/01 16:42:06 <0> process_requests[26719]: oprd received string 56 -1 0 0 DECTZ89CDEC0 *NULL* *NULL* 10/17/01 16:42:06 <0> process_requests[26719]: received TERMINATE request 10/17/01 16:42:06 <1> rdevmi[26642]: INITIATING 10/17/01 16:42:07 <0> rdevmi[26642]: got CONTINUE, connecting to ltid 10/17/01 16:42:07 <0> process_requests[26642]: oprd received string 56 -1 0 0 DECTZ89CDEC0 *NULL* *NULL* 10/17/01 16:42:07 <0> process_requests[26642]: received TERMINATE request The "daemon" debug log on the vmd/DA (volume database) host will show many repeated registration requests: 00:02:32 [237] <4> process_request: DA_REGISTER requested 00:02:32 [237] <2> da_register_shared_drives: server read: myhost 5 0 - DECTZ89CDEC0 DECTZ89CDEC1 DECTZ89CDEC2 DECTZ89CDEC3 DECTZ89CDEC4 DECTZ89CDEC5 DECTZ89CDEC6 00:02:32 [237] <2> da_register_shared_drives: returning scan host record for 7 drives registered by host myhost for host myhost 00:02:32 [237] <4> da_register_shared_drives: REGISTRATION of 7 drives for host myhost completed successfully Additional Notes: The race condition only occurs if servers are incorrectly configured, or if there are system and/or network problems. (NetBackup Servers) -------------------------------------------------------------------------------- Description: Device reset does not make a tape drive available if the drive is shared (multihosted), assigned to the host where the reset is performed, and there is no active mount request. Reset of a shared, unassigned drive does not make the drive available, even if the vmd/DA reservation is removed. Workaround: Stop and restart ltid, the Media Manager device daemon and NetBackup Device Manager service, on the host where the drive was assigned. (NetBackup Servers) -------------------------------------------------------------------------------- Description: Shared tape drives remain assigned indefinitely, even after unmounts are requested if the robot goes into the logical DOWN state while drives are assigned. A symptom of this problem are messages similar to the following appearing in the system log (or Event Viewer): Oct 16 16:05:49 myhost acsd[11984]: ACS(45) going to DOWN state, status: Timeout waiting for robotic command Oct 16 16:07:56 myhost acsd[11984]: ACS(45) going to UP state Oct 16 16:25:27 myhost acsd[11984]: Drive not set to IN USE prior to DISMOUNT Workaround: Stop and restart ltid, the Media Manager device daemon and NetBackup Device Manager service, on the host where the drive was assigned. Additional Notes: Drives will get left assigned only if there is a period of robotic control unavailability while drives are assigned, such as when: - There are network communication problems (with ACS, TLH, or TLM robots). - There are robotic hardware problems (with TL8 or TLD robots) including if the robot goes offline. - The robotic daemons/processes (acsd, tl8d, tldd, tlhd, tlmd) are stopped and restarted. (NetBackup Servers) -------------------------------------------------------------------------------- Description: In the 110539-08 patch with respect to "MEDIA ID BARCODE CHARS" the vmupdate executable cores on NT, if the vm.conf file does not exist. Workaround: If the vm.conf file exists, then this error would not occur. There need not be any specific entries in it, it just has to exist. (NetBackup Servers) -------------------------------------------------------------------------------- Description: The bpbkar process got a SIGSEGV and core dumped when a TIR backup with move detection was done with an exclude_list that excluded everything in the filelist from the class. (NetBackup Clients) -------------------------------------------------------------------------------- Description: In multiple-initiator (multiple-HBA) environments (such as SSO configurations), some form of device level protection is required to avoid unintended sharing of tape devices and possible data loss problems. The only widely available technique for this purpose is to use SCSI reserve/release functionality. NetBackup now uses SCSI reserve/release commands to improve data integrity in SSO (and also in non-SSO) configurations. SCSI reserve/release operates at the SCSI target level and depends on the fibre-to-scsi bridge or native fibre device hardware working correctly. This new functionality does not allow NetBackup to share tape devices with other applications. Prior to access from other applications, either use tpreq to mount media into the tape drive, or DOWN the drive. Additional Notes: *SCSI Reserve/Release Commands When a device receives a SCSI reserve command, it will no longer process most commands from any other HBA until the reserving HBA issues the SCSI release command. If an application sends a command to a reserved device, the device will fail the command by returning a status of RESERVATION CONFLICT. The only exceptions to this action are the Inquiry, Log Sense, Report LUNs, and Request Sense commands. A device stays reserved until one of the following occurs. The device is --Released by the HBA that reserved it. --Released by some sort of TARGET or LOGICAL UNIT RESET. These resets are protocol dependent, and differ between parallel SCSI and FCP (SCSI on fibre channel). These resets may be issued from any HBA. --Power-cycled. --Released by fibre channel LOGO/PLOGO/PRLI/PRLO/TPRLO or failed discovery (link actions). A negative effect of SCSI reserve can occur if the reserving HBA stops working (for example, due to a system crash or hardware failure). All devices reserved by the HBA stay reserved until the reservation is removed or broken. The reservation can only be removed by the original HBA, which means the system must be re-booted. In the case of a hardware failure, this is not possible. To break a reservation the device must be reset. This can be done by any of the following: --SCSI reset --Bus device reset --LUN device reset --Power cycle --Fibre channel link actions (may break reservations). SCSI reserve and SCSI release commands are mandatory for all SCSI-2 and SCSI-3 devices. See the SCSI-2 standard for a detailed description of SCSI reserve command operation and behavior. *How NetBackup Uses SCSI Reserve/Release Commands The following topics explain how NetBackup uses SCSI reserve/release commands in an SSO environment (or any other multiple-initiator environment). **Issuing the Reserve NetBackup processes (bptm, bprecover, and bpbackupdb) that read or write tape media issue a SCSI reserve command to the tape device that contains the media in use (during the open process). Once the reservation is established, all other HBAs are locked out of this tape device. This reservation prevents other HBAs from issuing commands that can cause data loss. This reservation does not prevent other applications from using the same device on the server with the reservation and causing data loss (for example, someone issueing a UNIX mt command). **Checking for Data Loss The bptm process detects data loss by reading the tape position and then checking the actual position against the expected position. If the actual position is less than the expected position (at the end of the backup process), the following will occur: --The tape is frozen. --The backup fails. --The following error message entry is placed in the error log: "FREEZING media id xxxxxx, External event caused rewind during write, all data on media is lost" If the SCSI reserve/release feature is not enabled on your servers, data loss can be caused by configuration errors, like incorrect paths, multiple master servers, incorrect SSO configurations and third-party or operating system utilities. If the SCSI reserve/release feature is enabled on all servers, then the cause could be third-party or operating system utilities running on the server that is also running the backup operation. Unfortunately data loss cannot be prevented, just recognized after the fact. The NetBackup catalog is not cleaned up to remove information on prior backup sessions lost. The bpexpdate command must be run on the media id to clean up the catalog. This position check may be disabled by creating a file "/usr/openv/netbackup/db/config/NO_POSITION_CHECK" (UNIX) **Checking for tape/drive configuration errors The bptm process detects configuration errors by reading the tape position and then checking the actual position against the expected position. Any configuration problem that causes the actual position to be greater than the expected position (at the end of the backup process), the following will occur: --The tape is frozen. --The backup fails. --The following error message entry is placed in the error log: "FREEZING media id xxxxxx, too many data blocks written, check tape/driver block size configuration" The source of the configuration problem needs to be identified and corrected. The most common configuration error is to fail to configure the driver for variable length blocks. A second source could be an error in the tape driver's configuration data. On Solaris, this could be in the /kernel/drv/st.conf. Please review the Media Manager Device Configuration Guide for the operating system you are using. The backup may be usable, in which case the image will need to be imported before restores can be done (bpimport). This position check may be disabled by creating a file "/usr/openv/netbackup/db/config/NO_POSITION_CHECK" (UNIX), or "INSTALL_PATH\netbackup\db\config\NO_POSITION_CHECK" (Windows NT/2000). **Issuing the Release After a NetBackup process is done with the media, a SCSI release is issued as part of the unmount operation. This release frees the device for access by another HBA. In addition, at the beginning of the startup process, avrd issues a SCSI release to all configured tape devices that are currently in the UP state. This is done to release devices that were reserved at the time of a system re-boot or crash. The SCSI release command will return tape devices to general availability after a system crash. **Error Recovery To recover a device that is reserved by an HBA that crashes or was otherwise unable to issue the SCSI release command, you can use the following option for the Media Manager vmoprcmd command: vmoprcmd - releasebyname drive_name. This option causes the host to attempt to release the drive (using the SCSI release command). See your Media Manager System Administrator's Guide for the complete syntax and more information on using the vmoprcmd command. *SCSI Reserve/Release Logging and Conflict Notification The bptm process logs all SCSI reserve/release commands. The bptm log should be checked on all hosts to ensure the SCSI reserve operation is being logged (look for "SCSI RESERVE"). The avrd process monitors all tape devices. NetBackup manages access to tape devices, such that a properly configured system will not receive the RESERVATION CONFLICT status from a tape device. If the avrd process gets a RESERVATION CONFLICT status, avrd changes the status of the device to PEND and logs this event in the syslog. The log message will be "Reservation Conflict status from DRIVENAME (device NUMBER)". When the conflict is resolved, the message in the log will be "Reservation Conflict status cleared from DRIVENAME (device NUMBER)". If this conflict occurs, some sort of mis-configuration is present (the tape is reserved, but should not be) and the configuration problem should be corrected. A possible cause of this conflict is if an operating system crashes or a hardware failure has left a device reserved (see "Issuing the Release"). Also in the Device Monitor or vmoprcmd output, PEND in the Control column means that a reservation conflict has occurred. *Issuing Manual Resets to Break a Reservation On the following UNIX operating systems, you can try to reset a reservation conflict by using the following reset commands. CAUTION: The reset operation may reset other devices in your configuration. Loss of data is also possible. Alternate methods of breaking the reservation on a device (using switch and bridge hardware) should be tried first. **Sun Solaris Issue the following commands: 1. mt -f drive_path_name forcereserve 2. mt -f drive_path_name release See the mt(1) man page for more information. *Controlling SCSI Reserve/Release In this release using SCSI reserve for data integrity is disabled by default, but can be enabled by creating the file /netbackup/db/config/ENABLE_SCSI_RESERVE. If the config directory doesn't exist, create it. *SCSI Reserve/Release Requirements and Limitations The following topics cover important requirements and restrictions. **Requirements The following requirements are needed --There must be passthru driver access to all shared drives. The passthru driver must be installed and all required paths created. See the NetBackup Media Manager Device Configuration Guide for information on configuring and using the passthru driver for various UNIX operating systems. --Host operating systems must be set up properly to enable the SCSI reservation capability. --Users of Sun Solaris 2.6 or 7 must install a st driver patch to avoid a problem that keeps the device reserved when it should not be. For Solaris 2.6, the minimum patch level required is 105847-06. For Solaris 7, the minimum patch level required is 107460-06. **Limitations This VERITAS implementation using SCSI reserve/release has the following limitations: --SCSI reserve/release is not applicable for NDMP configurations (no reserve command is available). --Cluster environments or multi-path environments with fail-over capability may leave devices reserved. If fail-over doesn't break device reservations, then the NetBackup SCSI reserve/release function must be disabled. --Cluster environments or multi-path environments with dynamic path sharing (TRU64 for example) will cause backup and restore failures if the path changes during these operations. If path sharing cannot be eliminated, then the NetBackup SCSI reserve/release function must be disabled. *********************************************************** The Tape Device Monitor (dm_stape) of the Event Monitoring System (EMS) should be configured to not run on hosts participating in the tape backup SAN. Part of the EMS service periodically polls the tape devices to monitor their health. When this occurs from one server, while another server is using the tape device, it can interfere backup operations, causing those operations to timeout and abort. Currently there are two ways to disable the dm_stape monitor while keeping the Event Monitoring System enabled. If a customer wishes to completely disable event monitoring it can be done by running the tool /etc/opt/resmon/lbin/monconfig and selecting '(D)isable Monitoring'. Preventing dm_stape from polling If the "POLL_INTERVAL" value in the /var/stm/config/tools/monitor/dm_stape.cfg file is set to zero, the monitor will still run, but it will not log any events or poll the devices (i.e. it will not send any SCSI commands). Due to a bug in all IPR releases before IPR0109, this will not work properly. This bug has been partially fixed in the IPR0106 release; it will be completely fixed in the IPR0109 release. For the IPR0103 release, there are 2 patches available (which also partially fix the problem) as follows: After either of these patches is installed, the user needs to manually set the "POLL_INTERVAL" value in the /var/stm/config/tools/monitor/dm_stape.cfg file to zero to stop the monitor from polling. IMPORTANT NOTE: For the IPR0106 release and for the IPR0103 release with either patch, the dialoged process must be running when you set the POLL_INTERVAL value to zero. Otherwise, the monitor will fill the api.log file with error messages (until the hard disk space is used up) and consume most of the CPU time. Under no circumstances should dialoged or the STM diagnostics be shut down! *********************************************************** (NetBackup Servers) -------------------------------------------------------------------------------- Description: When SSO is used in an environment where (local area) network failures occur (for example, one with faulty network interface cards), the remote drive retry logic does not work, and related system log messages do not specify the kind of error that has occurred. The following errors would be seen in the system log: "Remote drive scan failed for multihosted drive, daemon terminated" and "Remote scan failed on host XXX, drive YYY, Unknown Error (4)." The first of these errors will no longer occur because the retry logic has been corrected. The second error has been changed to log the error "Remote scan failed on host XXX, drive YYY, Interrupted system call". Workaround: Correct the faulty network. (NetBackup Servers) -------------------------------------------------------------------------------- Description: Dumplevel for incremental backups would be set incorrectly for NDMP causing more files to be included in the backup than should be. (NetBackup Servers) -------------------------------------------------------------------------------- Description: For a client machine with: - A client with multiple NICs. - The hostname of the client is "XYZ". - The hostname of the primary NIC is also "XYZ". - The CLIENT_NAME of the client is "XYZ-BU". - The hostname of the NIC connected to the backup network is also "YXZ-BU". With a scenario such as this, it is possible for bprestore and other client commands to pass the incorrect hostname on requests to the master server that will cause user restore and other requests to fail. Workaround: Use the REQUIRED_INTERFACE configuration entry on the client. (NetBackup Clients) -------------------------------------------------------------------------------- Description: During a NetBackup DB backup on an AUSPEX master server, the backup fails because it could not find some libraries. This only happens if /var/opt/lanman/regfiles/reg.ini exists. The error report will have the following error: ERR - /opt/lanman/lib/liblmxaclapi.so NetLink library is missing. NetLink ACLs cannot be backed up. Workaround: Move the /var/opt/lanman/regfiles/reg.ini somewhere else during the NBU DB backup. (NetBackup Clients) -------------------------------------------------------------------------------- Description: This is to add support for HCART drive types in 3.4 BusinesServer. Additional Notes: HCART drivetypes are configured either in RSM libraries or tld libraries. NetBackup BusinesServer does not support the configuration of multimedia libraries. In other words, you may not configure a library with both an hcart drive and a dlt drive. (NetBackup Servers) -------------------------------------------------------------------------------- Description: A standalone tape drive, configured for both SSO and DO_NOT_EJECT_STANDALONE, will not complete a backup that spans media. The full media is ejected, but the next media is not recognized when it is loaded into the drive. (NetBackup Servers) -------------------------------------------------------------------------------- Description: Recognize Solaris 9 (and Solaris_x86_9 if applicable) as valid client types. Additional Notes: NetBackup for EMC not supported on Solaris 9. Also, Extended File Attributes are not supported with the initial release of NetBackup 4.5. Support for this on UFS, NFS & TMPFS file systems will be added in a later release. This feature is a late addition to Solaris and is the reason that NetBackup cannot support it this release. You can monitor NetBackup patch READMEs to learn when NetBackup will support the Extended File Attributes feature. (NetBackup Servers) -------------------------------------------------------------------------------- Description: When trying to use set UID to root binaries with VOPIE authentication, it is likely that the message "authentication failed (160)" will appear when trying to use one of those binaries as a non-root user. Set UID to root binaries are typically created by the nonroot_admin script. Additional Notes: Unable to use set UID to root binaries with VOPIE authentication on those platforms. (NetBackup Clients) -------------------------------------------------------------------------------- Description: VOPIE authentication failures may occur when the case of host names differ between NetBackup machines. For example, if a NetBackup server thinks that the host name of a client is "abc" and the client thinks its host name is "ABC", authentication errors (error code 160) are likely to occur between the client and server if VOPIE authentication is set up between the two machines. Workaround: Make sure the case of host names is consistant on all machines. (NetBackup Clients) -------------------------------------------------------------------------------- Description: When doing TIR restores from multistream backups created with ALL_LOCAL_DRIVES, the filesystem mountpoints from the root filesystem backup get displayed in the GUIs so when they are selected, you only get the mountpoint restored instead of the full filesystem. Workaround: Select the filesystem with a slightly earlier time stamp. (NetBackup Clients) -------------------------------------------------------------------------------- Description: When multiple jobs failed which were started from the same class to multiple clients, several problems could occur. The status returned to the requesting scheduling process may return with a 213 instead of the real status. This may lead to incorrect handling of the resource_busy (134) status. This also would sometimes cause the status to be sent more than once, which could eventually result in filling the request message queue. (NetBackup Servers) -------------------------------------------------------------------------------- Description: If you specify a progress log on the >bpbackup command for an immediate backup, the server will try to log data to the client. If bprd can't log to the client (perhaps because of a bad path name), the backup is terminated. In older releases this did not happen. For immediate backups, bprd will now turn off progress logging if the requesting host is not the same as the target host. (NetBackup Servers) -------------------------------------------------------------------------------- Description: If the start and end times of a job are identical, the elapsed time as delivered to the java activity monitor is incorrect. It should be exactly 0 seconds. (NetBackup Servers) -------------------------------------------------------------------------------- Description: When formating optical disks on the Solaris 8 6/00 Release or later, the following tpformat errors may occur: ERROR 1: #./tpformat -ev s5121a -d odiskwm tpformat: Cannot open /tmp/s5121a, Device busy The Solaris 8 6/00 Release introduced changes to the Solaris volume manager daemon (vold), which attempts to manage all removable media devices. When vold manages an optical disk, NetBackup cannot access it. To resolve this issue, edit /etc/vold.conf to comment out the following line as shown below. After this change, vold will no longer attempt to manage optical disks. #use rmdisk drive /dev/rdsk/c*s2 dev_rmdisk.so rmdisk%d ERROR 2: #./tpformat -ev s5121a -d odiskwm No disks found! The Solaris 8 6/00 Release introduced a change to the Solaris format command. The "e" option is now required to allow access to removable media devices. The tpformat command, which uses the Solaris format command internally, now specifies the "e" option. (NetBackup Servers) -------------------------------------------------------------------------------- Description: The Device Configuration Wizard can now configure drives in the following robotic libraries: DELL PV136T, HP A4845A, HP A4846A, and HP A4853A. (NetBackup Servers) -------------------------------------------------------------------------------- Description: Large ndmp backups were failing when the write in flush_imebuf got interrupted by a signal, causing the data transferred to be less than the entire write. The log entry looks like this: 20:38:51 [25844] <32> bpfsmap: flush_imebuf: write to tmp inomap file failed errno=0 20:38:51 [25844] <2> write_data_ndmp: bpfsmap longjmp: setjmp returned 14 (NetBackup Servers) -------------------------------------------------------------------------------- Description: The NetBackup process (bpbkar32) on NT and W2K NetBackup clients waits for a status message from the NetBackup server at the end of the backup job. This status is used to determine if the archive bits should be cleared for the files that were backed up (among other things). Essentially, the archive bits will be cleared only if this message indicates a successful backup. If there is a delay on the server before sending this status message to client, the socket connection could time out and the subsequent send of the status message would fail. Since the archive bits are not being properly cleared in this case, one end result of the problem is that incremental backups back up more data than needed or intended. There may be an entry such as the following in the bpbkar debug log: 4:54:21.209 PM: [1000.1592] <2> tar_base::V_vTarMsgW: WRN - timeout (300) while waiting for message: 'Clear Archive Bits' Note that the timeout value (300 in this example) is a configurable parameter, so this parameter may need to be increased to a larger value on the client if this condition is encountered. However, prior to this fix, increasing this parameter would not work. The status message would still fail to be seen by the client. (NetBackup Servers) -------------------------------------------------------------------------------- Description: Enhancement to implement -mheader option on bpmedialist command. This is similar to -mcontents, but only reads and prints the media header. The backup headers are not read or printed. (NetBackup Servers) -------------------------------------------------------------------------------- Description: When creating a rename file on a Win2k machine (e.g. using notepad), carriage return (\r) character may be created. So when restoring a standard UNIX file, the \r is appended to the end of the file name that is restored. The customer had been running into this when trying to do an AFS restore from a Win2K server to a Solaris client using Netbackup 3.4. AFS does not like the \r character and the restore fails. (NetBackup Clients) -------------------------------------------------------------------------------- Description: In certain SSO configurations where the device hosts had very long host names, vmd would core dump or generate access violations. (NetBackup Servers) -------------------------------------------------------------------------------- Description: In an SSO configuration that encounters errors, the memory usage of the device management service/daemon (ltid) will continue to grow. (NetBackup Servers) -------------------------------------------------------------------------------- Description: Update the external_drive_mappings/external_robot_mappings to the latest available. Added sony AIT 3 drives. Added additional inquiry string for the Quantum Super Dlt drive Added support for Overland NEO and NEC Dlt Autoloader. Added support for IBM 3584L42 Added support for Exabyte 215 as tl8. Add ATL P4000 and ATL P7000 libraries. Add support for DELL PV-136T Added several ADIC and HP optical libraries You may want to check the VERITAS web site for a later version of these files. (NetBackup Servers) -------------------------------------------------------------------------------- Description: When circular logging is enabled, if a cumulative incremental or differential incremental backup is requested, only those logged databases that have had their DBIID changed since the last full (or in the case of a differential incremental, last full or incremental backup) will be backed up during the incremental backup. This is correct when archive-style logging is enabled, because the incremental backup relies on the backup of the available transaction log extents as the incremental. The only time the database needs to be backed up in this scenario is when the DBIID has changed and we need an initial backup of the database (with the new DBIID) to allow a later database recovery using the transaction from the log extents. In the case of circular logging, the transaction log extents are NOT backed up. We cannot rely on the transaction log extents to be the incremental backup. In order to incrementally protect your databases ( in case of a loss of the transaction log extents), all databases that have been modified since the last full (or in incremental for a differential incremental) should be backed up. (NetBackup Clients: Solaris2.6) -------------------------------------------------------------------------------- Description: Special patch for LTO support on BusinesServer (patch 948) prevented the Volume Configuration wizard from configuring cleaning tapes. Workaround: Manually add cleaning tapes using the volumes/add menu option or vmadd command line. (NetBackup Servers) -------------------------------------------------------------------------------- Description: Unable to correctly configure cleaning tapes for TLD robots that contain HCART drives for BusinesServer machines via the Volume Configuration Wizard. Workaround: Use the robot inventory dialog in Media and Device Management to configure cleaning tapes for TLD robots that contain HCART drives on BusinesServer machines. (NetBackup Servers) -------------------------------------------------------------------------------- Description: On active filesystems, the directory block in the kernel space may get rearranged and cause a backup that is reading the same directory block to get an incomplete list of the files in the directory. This will cause the backup to be incomplete. Workaround: Do a TIR backup with move detection to read the directory as fast as possible and avoid restructuring problems. (NetBackup Clients) -------------------------------------------------------------------------------- Description: At Oracle9i, temporary tablespaces cannot be put into backup mode. The select statement in the bpstart_notify.oracle_bli script includes the temporary tablespace. This causes any BLI backup which performs a alter tablespace to fail. (NetBackup Clients: Solaris2.6) -------------------------------------------------------------------------------- Description: A bpbackup -W command will return a 134 error unders certain conditions. This happens when the backup is a user type, the failure count has already exceeded the configured limit, and a 134 (resources busy) condition occurs. The status is returned to the bpbackup command but the job requeues and waits for resources. The status should not be returned. (NetBackup Servers) -------------------------------------------------------------------------------- Description: When opening the JAVA Activity Monitor on the Masterserver, no jobs are shown and the error "unexpected message received" is reported. This is a reproducible anomaly on Tru64 machines.. and alleged on other Unix platforms. The bpjava-[s]usvc log will show an entry such as: >> 11:35:14 [295315] <2> job_output: tag = 10 = EXEC_BPDBJOBS >> 11:35:14 [295315] <2> job_output: EXEC_BPDBJOBS st_size = 0 >> 11:35:14 [295315] <2> job_output: EXEC_BPDBJOBS read pipe is empty >> 11:35:14 [295315] <16> job_output: EXEC_BPDBJOBS line count is NULL!!! when/while the bpdbjobs process is still active/alive (NetBackup Servers) -------------------------------------------------------------------------------- Description: Lotus Notes restores and Single Instance Store and Hard Link restores may fail in certain configurations. The Single Instance Store and Hard Link restore problems are limited to the Windows platform. The problem is most readily seen when performing a user directed restore. During the course of the restore, the user may end up seeing like this in the progress log: 09:26:01 (37.001) INF - TAR BEGIN SECONDARY RESTORE 1 -1 00 < LINES OMITTED - THESE WILL REFER TO FILES THAT NEED TO BE > < RESTORED VIA A SECONDARY RESTORE > 09:26:01 (37.001) INF - TAR END SECONDARY RESTORE However, there is never any indication that the files to be restored via the secondary restore are ever restored. For Lotus Notes, the restore will never complete. It will hang until those files are restored. For Single Instance Store and Hard Links, the original files restores will more than likely be empty files or refer to a file that is not there. (NetBackup Servers) ================================================================================ Description: Unable to correctly configure cleaning tapes for TLD robots that contain HCART drives for BusinesServer machines via the Volume Configuration Wizard. Workaround: Use the robot inventory dialog in Media and Device Management to configure cleaning tapes for TLD robots that contain HCART drives on BusinesServer machines. (NetBackup Servers) ================================================================================ Description: When the Japanese version of NetBackup Java is run in the "ja" locale, the Activity Monitor help screens display in English rather than Japanese. (NetBackup Clients) Description: A denial of service attack, as reported at http://www.securityfocus.com/archive/1/175116 N+1 connections to the bpjava-msvc port (13722) via inetd, where N is the number of random ports available, can result in exhausting the number of ports available and a denial of service. The N+1st connection must be made within 1 minute of the first, as the first would timeout, thus creating an available port. Note that no request is made to bpjava-msvc, only the connection established. Additional Notes: Do not open a 'port' on the master server for communication with a child user server when the master is transient (currently Unix only). This unnecessary port/acquisition can be the target of a denial of service attack. When transient, a new bpjava-msvc is created, via inetd, for every connection to 13722. It takes 60 seconds for bpjava-msvc to timeout if no valid request is sent following the connection, more than enough time to attempt multiple bpjava-msvc connections, and consume multiple (all) ports needlessly for communication with a user server which will never be created. Even if the transient master did receive a valid request and was able to create a (valid) user server, the information sent back to the master is only useful IF the master is NOT transient. (NetBackup Clients) Description: System-created hard links were not properly restored using the Java interface, but would work via the Motif interface. If the hard links were user created, then both GUIs worked as designed. (NetBackup Clients) Description: Adds recognition of ACSLS 5.4 and 6.0 media and drive types. The Java and Windows NT/2000 Media and Device Management interfaces and the UNIX media management menu interface have been modified to recognize and configure media type mappings for the media types listed below. The Java Media and Device Management interface has also been modified to add ACS drives with a maximum drive index of 19. Upon initialization, the ACS robotic daemon/process will now recognize all drive types supported in ACSLS 5.4 and 6.0. ACS Default Media Manager Allowable Media Types Media Type Media Type through Mappings ---------- --------------------- --------------------- STK2P HCART2 HCART, HCART2, HCART3 STK2W HC2_CLN HC_CLN, HC2_CLN, HC3_CLN KLABEL HCART HCART, HCART2, HCART3 LTO_100G HCART HCART, HCART2, HCART3 LTO_50GB HCART HCART, HCART2, HCART3 LTO_35GB HCART HCART, HCART2, HCART3 LTO_10GB HCART HCART, HCART2, HCART3 LTO_CLN2 HC_CLN HC_CLN, HC2_CLN, HC3_CLN LTO_CLN3 HC_CLN HC_CLN, HC2_CLN, HC3_CLN LTO_CLN1 HC_CLN HC_CLN, HC2_CLN, HC3_CLN SDLT DLT3 DLT_CLN, DLT2_CLN, DLT3_CLN Workaround: Unrecognized ACSLS 5.4 and 6.0 media types will be returned as "UNKNOWN". These can be mapped to Media Manager media types in any of the media management interfaces as long as all unknown media are mapped to a single Media Manager media type. Additional Notes: New vm.conf Options for Use if Defaults Need to be Overridden ------------------------------------------------------------- ACS_STK2P = ACS_STK2W = ACS_KLABEL = ACS_LTO_100G = ACS_LTO_50GB = ACS_LTO_35GB = ACS_LTO_10GB = ACS_LTO_CLN2 = ACS_LTO_CLN3 = ACS_LTO_CLN1 = ACS_SDLT = where is one of the allowable media types. (NetBackup Clients) ================================================================================ ============= 110539-10 ============= Description: Fixes problem when restoring directories where setgid bits are used. Child directories inherited the setgid bit setting if the parent had it set, regardless of the directory's original setting when backed up. (NetBackup Clients) Description: A denial of service attack, as reported at http://www.securityfocus.com/archive/1/175116 N+1 connections to the bpjava-msvc port (13722) via inetd, where N is the number of random ports available, can result in exhausting the number of ports available and a denial of service. The N+1st connection must be made within 1 minute of the first, as the first would timeout, thus creating an available port. Note that no request is made to bpjava-msvc, only the connection established. Additional Notes: Do not open a 'port' on the master server for communication with a child user server when the master is transient (currently Unix only). This unnecesary port/acquisition can be the target of a denial of service attack. When transient, a new bpjava-msvc is created, via inetd, for every connection to 13722. It takes 60 seconds for bpjava-msvc to timeout if no valid request is sent following the connection, more than enough time to attempt multiple bpjava-msvc connections, and consume multiple (all) ports needlessly for communication with a user server which will never be created. Even if the transient master did receive a valid request and was able to create a (valid) user server, the information sent back to the master is only useful IF the master is NOT transient, and therefore its useless when the master IS transient. Resolves DDTS RSVmn15352. (NetBackup Servers) Description: BLIB incremental backups of large filesystems where little has changed take longer than before. (NetBackup Clients) Description: Exceeding the limit of 940 drives in an STK Library Station configuration. This fix increases the limit to 1680. The external symptom to this problem is acsd not starting. The system log file contains the following error message: Apr 5 16:23:35 bison acsd[17118]: DecodeQuery() Actual status: Process killed by signal Apr 5 16:23:35 bison acsd[17118]: ACS(1) unavailable: initialization failed: Process killed by signal (NetBackup Servers) Description: Restore failures would still have a final message of "the requested operation was successfully completed". (NetBackup Servers) Description: Using Compaq ESL9000 series robots (ATL P3000 rebadge). When they upgraded from 3.2 to 3.4, they began having problems ejecting tapes from the libraries. The third slot in the CAP seems to be know as a PTM (Pass Through Mechanism) slot, which is designed to allow libraries to be hooked together and pass tapes between themselves. An Onsite Compaq SE has stated that the PTM has never been activated. So now the operators can only eject two tapes, open the CAP pull them out, close the CAP, eject two more tapes, etc. DEC TL810s and DEC TL820s existed on this same media server. The TL820s DO have a PTM. Workaround: Workaround is to have all DEC TL820s or any other ATL 2640 or ATL 2640 rebadge on their own media server. In other words, this robot, the ATL 2640, does not play nice with other robots. (NetBackup Servers) Description: ACS partial robot inventory fails when encountering unknown ACS media types. The inventory operation fails with an error such as: Unrecognized ACS media type ordinal 17 invalid media type (9) For this to have occurred, partial robot inventory would have been enabled using a INVENTORY_FILTER configuration entry in the Media Manager configuration file, vm.conf. Workaround: Avoid using media types that are not recognized. The recognized media types are those available in ACSLS 6.0 and earlier, which include: 3480, 3490E, DD3A, DD3B, DD3C, DD3D, DLTIII, DLTIV, DLTIIIXT, STK1R, STK1U, EECART, JLABEL, STK2P, STK2W, KLABEL, LTO_100G, LTO_50GB, LTO_35GB, LTO_10GB, LTO_CLN2, LTO_CLN3, LTO_CLN1, and SDLT. If other media types are required and ACSLS is the server type, use full robot inventory in place of partial robot inventory. Otherwise, perform media management operations using the manual add/change volume interface instead of robot inventory update. (NetBackup Servers) Description: Currently, the last modification date of the Lotus transaction log extent is being checked to determine if the transaction log extent should be backed up as part of a cumulative incremental or differential incremental backup. Instead, when the BACKUP_TRANSACTION_LOGS file list directive is specified, all available transaction log extents should be backed up as part of the incremental, without first checking the last modification date of the transaction log extent. (NetBackup Clients) Description: If a colon ":" character was used in the name of a storage unit, drives that should have been available, might be treated as if they were down. Workaround: Remove all colons from storage unit names. (NetBackup Servers) Description: When backing up Lotus databases, a misleading warning message is displayed in the bpbkar log for each database that is backed up. An example of the warning message is: "WRN - Short read at byte 524288. Read 40488 bytes when attempting to read 524288, in file /db/notesdata/admin4.ntf.". (NetBackup Clients) Description: If a media mount request fails for a backup because the media is write protected or unmountable because of a media problem, NetBackup will now immediately FREEZE the media so it isn't requested over and over during the duration of a backup window. (NetBackup Servers) Description: The image files in the NetBackup image database do not get removed when Oracle images are expired through RMAN. (NetBackup Servers) Description: Adds support to NetBackup for AFS volumes larger than 2GB. This is an AFS 3.6 feature. (NetBackup Servers) Description: NDMP attached STK L700 robots intermittently crash tldtest/ tsdtest/tl8test when an "s s" command is entered. tldcd/ tsdd/tl8cd intermittently crash while doing inventory. The problem could occur with other library types with a large number of slots when used with NDMP. (NetBackup Servers) Description: Non-robotic volumes may cause vmdb_compact to fail. Log messages may look like: 07:35:25 [18351] <4> process_request: REINSERT requested 07:35:25 [18351] <2> insert_entries: server read: 1 BIRK01 ------ 11 -------- -------- 0 -1 NONE --- 0 0 0 0 0 root root 0 None - 994335039 0 0 0 0 0 0 0 0 - 0 0 34 0 0 0 0 0 - - 0 0 0 0 0 0 0 0 0 0 0 ------ 07:35:25 [18351] <2> string_to_record: using server version: 34 07:35:25 [18351] <16> server_validate_dbrec: invalid robot number -1 07:35:25 [18351] <16> insert_entries: Invalid record received from client: Permission denied (13) 07:35:25 [18351] <16> insert_entries: insert failed -- 0 entries inserted 07:35:25 [18351] <16> insert_entries: INSERT FAILED (stat=0)-- 0 entries inserted 07:35:25 [18351] <16> listen_loop: abnormal exit: 13 in particular the line: 07:35:25 [18351] <16> server_validate_dbrec: invalid robot number -1 (NetBackup Servers) Description: The reinsert volume operation fails with records that may be entirely valid. The reinsert volume operation is used by various media management operations including the command line volume database compaction utility, vmdb_compact. The symptom of this issue is when the volmgr/debug/daemon log file contains lines such as the following: 08:21:59 [18522] <4> process_request: REINSERT requested 08:21:59 [18522] <2> insert_entries: server read: [[ volume record here ]] 08:21:59 [18522] <2> string_to_record: using server version: 34 08:21:59 [18522] <16> insert_entries: INSERT FAILED (stat=5)-- 0 entries inserted 08:21:59 [18522] <16> listen_loop: abnormal exit: 5 The status associated with the INSERT FAILED line could be anything. The key is that failure occurs on the REINSERT volume request. (NetBackup Servers) Description: Hard links that happen to be of length 100, do not get restored. Titan Case 150-053-467. (NetBackup Clients) Description: If hardlink names are exactly 100 characters long, the stored linkname is non-null terminated. This causes the next field to be concatenated with the linkname. Restores would not work in this case. tar was changed so it does not look for a null termination. (NetBackup Clients) Description: The problem happens during backups on Windows NT/Windows 2000. While backing up files, the OS may report an error while reading the file. This causes an incomplete/invalid file to be present in the image. Backups of the 'outlook.pst' file has displayed this behavior when Outlook is running. When browsing for restores, everything looks ok. The file appears in the catalog and the displayed size is correct. The real problem actually happens during the restore as a file of incorrect size gets restored. With the example of 'outlook.pst', it will end up getting restored as a 0 byte file. The backup side has been changed to detect this type of problem and keep incomplete/invalid files from appearing in the NetBackup database. This will prevent the files from being restored. (NetBackup Clients) Description: First release of NetBackup Configuration Validation Utility (NCVU) for NetBackup 3.2 and 3.4 on Solaris, AIX and HP-UX. (NetBackup Servers) Description: For some DHCP clients (including Windows 2000 and Mac OS X), the Dynamic Addressing configuration described in the System Administration Guide would not work. The problem is that the clients wouldn't (or couldn't) pass a host name on bprd requests that the master server would accept. This would case failed bpdynamicclient, bpbackup, bpclimagelist, bplist, and bprestore commands from the client. In addition, scheduled backups by the server would also fail. In addition non-DHCP clients can have problems with user backups and restores (bpbackup, bpclimagelist, bplist, bprestore) if the client and server are in different network domains. Workaround: For the non-DHCP problem, you have to make sure that the client name is specified as a fully qualified host name in the class definition on the master server and in the "CLIENT_NAME" configuration entry on the client. Additional Notes: (NetBackup Servers) Description: The case where a restore was requested where the files in the file list were not found was returning a success code. A flag was added to catch this case and return error code 5 ("the restore failed to recover the requested files"). (NetBackup Servers) Description: A backup is run repeatedly. When the backup is done, it still appears due and starts again the next time the scheduler wakes up. This happens when a class that has multiple data streams available, multiple schedules of the same type(ex full) and only one data stream is generated. Workaround: Uncheck multiple data streams in the class. (NetBackup Servers) Description: Some Fibre Channel tape drives can report status "QUEUE-FULL", which can cause the avrd process to rewind the media. avrd process will now ignore "QUEUE-FULL" status. (NetBackup Servers) Description: The Device Configuration Wizard can now detect and configure Qualstar libraries. (NetBackup Servers) Description: Added the ability to backup and restore NT Access Control List (ACL) and Attribute information for files stored on the VERITAS ServPoint NAS Appliance. (NetBackup Servers) Description: If new jobs are queued to scheduler while waiting for completion of dbbackup, a new start session notify will not be executed. (NetBackup Servers) Description: tapes get left mounted on sn6000 if the drives are not configured as multihosted. Workaround: Drives must be configured as multihosted and max drives this master set to non-zero. Additional Notes: The change eliminates the need to configure the drives as multihosted. The sso license must still be installed and max drives this master must be non zero. (NetBackup Servers) Description: For Solaris 8 clients, the bpcoverage File System Backup Coverage Report incorrectly identifies "mvfs" as true local file system. This report should ignore mvfs file systems. Note that the check_coverage script invokes bpcoverage. (NetBackup clients: solaris8) Description: The Device Configuration Wizard can now configure STK L180, HP A5617A, and HP A5597A libraries which are running a SCSI-3 compliant firmware revision. The drives are now properly configured within the library, instead of being configured as standalone drives. Workaround: The STK L180, HP A5617A, and HP A5597A libraries and drives can be configured manually using any of the available device configuration GUI or command interfaces. (NetBackup Servers) Description: The wrong error status is returned to bpbackup. This occurs when multiple statuses are available due to multiple clients in the class and a client not being specified and more than 1 backup fails. If more than 1 failure occurs, only one status can be returned and it is unknown which it will be. This behavior has been modified to return the highest numeric status. (NetBackup Servers) Description: Quantum SDLT drives will perform poorly during backup operations without firmware 35 or above and this patch. This patch will send a vendor unique form of the Load command to SDLT drives to avoid long delays during WEOF (write end of file mark) ioctl processing. Driver timeouts are possible, as the WEOF command can typically take up to six minutes (13 minutes worst case) without both SDLT firmware 35 and this patch. (NetBackup Servers) Description: FlashBackup restores fail with an assert in sfr.c after patch P0850562 is installed. (NetBackup Servers) Description: Adds recognition of ACSLS 5.4 and 6.0 media and drive types. The Java and Windows NT/2000 Media and Device Management interfaces and the UNIX media management menu interface have been modified to recognize and configure media type mappings for the media types listed below. The Java Media and Device Management interface has also been modified to add ACS drives with a maximum drive index of 19. Upon initialization, the ACS robotic daemon/process will now recognize all drive types supported in ACSLS 5.4 and 6.0. ACS Default Media Manager Allowable Media Types Media Type Media Type through Mappings ---------- --------------------- --------------------- STK2P HCART2 HCART, HCART2, HCART3 STK2W HC2_CLN HC_CLN, HC2_CLN, HC3_CLN KLABEL HCART HCART, HCART2, HCART3 LTO_100G HCART HCART, HCART2, HCART3 LTO_50GB HCART HCART, HCART2, HCART3 LTO_35GB HCART HCART, HCART2, HCART3 LTO_10GB HCART HCART, HCART2, HCART3 LTO_CLN2 HC_CLN HC_CLN, HC2_CLN, HC3_CLN LTO_CLN3 HC_CLN HC_CLN, HC2_CLN, HC3_CLN LTO_CLN1 HC_CLN HC_CLN, HC2_CLN, HC3_CLN SDLT DLT3 DLT_CLN, DLT2_CLN, DLT3_CLN Workaround: Unrecognized ACSLS 5.4 and 6.0 media types will be returned as "UNKNOWN". These can be mapped to Media Manager media types in any of the media management interfaces as long as all unknown media are mapped to a single Media Manager media type. Additional Notes: New vm.conf Options for Use if Defaults Need to be Overridden ------------------------------------------------------------- ACS_STK2P = ACS_STK2W = ACS_KLABEL = ACS_LTO_100G = ACS_LTO_50GB = ACS_LTO_35GB = ACS_LTO_10GB = ACS_LTO_CLN2 = ACS_LTO_CLN3 = ACS_LTO_CLN1 = ACS_SDLT = where is one of the allowable media types. (NetBackup Servers) Description: The vmupdate command could core dump or cause an exception condition during a robot inventory update operation if patch NB_34_1 is installed. (NetBackup Servers) Description: The problem would cause the output to the screen to be wrong, after media is added to a robot. This would only occur if the feature "Media ID barcode chars" was being used. The media IDs that would be created would be correct, but the output to the screen would not. (NetBackup Servers) Description: Enhancement to the tape cleaning support: On tape dismount, the CLEAN_NOW and CLEAN_PERIODIC TapeAlert flags are checked. If either of these flags are set, an automatic tape cleaning request is queued. Additional Notes: Quantum DLT7000 and Quantum DLT8000 will now work with NetBackup's implementation of TapeAlert. (NetBackup Servers) Description: A marginal DLT tape may take longer to mount than the 5 minute timeout allowed by NetBackup. The following message will be written to the system log for mounts that exceed the timeout: tldd[5871]: TLD(0) [5871] waited 60 times for ready, drive 1 The mount timeout for DLT drives has been increased to 10 minutes. (NetBackup Servers) Description: An internal routine was performing control character substitution on the entire result string instead of on the static (i.e. format) string retrieved from the message file. This causes problems with strings such as D:\temp that are inserted into the result string at run time. NOTE, if the format string is not found in a message file (in the normal NetBackup English installation, for example), the control character substitution is not performed and this error does not occur. (NetBackup Servers) Description: The bpsched log entry add_to_drvstu_list has been removed. This message is not useful for debug. Also messages for host_equal and get_stunit_primary_hname have been changed so they log only when verbose is greater than 3. (NetBackup Servers) Description: The NetBackup debug logs cannot log more than 2 gigabytes of information. This may only occur when the logs are linked to another filesystem. (NetBackup Servers) Description: System log messages are repeated several dozen times in the system log in some cases when the requested media for a robotic mount request is found in another robotic tape or optical drive. If the robotic daemon/ process is running in verbose mode, a message such as the following may be found in the system event log: tlmd[12525]: DecodeMount(): TLM(1) drive 15, Actual status: Requested tape in other or non-configured drive above message repeats 213 times Note: the second message "above message repeats 213 times" is the symptom of the problem. It is normal expected behavior for the message "Requested tape in other or non-configured drive" to be logged. (NetBackup Servers) Description: Adding Oracle 9i support in the bli scripts and corrects when bpstart_notify.xxx.bli fails, tablespaces are left in back mode and bpstart_notify.xxx.bli script fails when encountering a Readonly tablespace. (NetBackup clients: solaris2.6) Description: Corrected a problem of extra white spaces when setting up ORACLE_LOG_LIST. Added an "if" statement to differentiate between and empty list and a non-empty list. (NetBackup clients: solaris2.6) Description: If a client is added to the client database without a hostname/ip address (typical for non-DHCP clients), bplist (and other) requests from the client will get error code 133 ("invalid request") from bprd. Workaround: Use the bpclient admincmd to add a hostname to the client database. For example, for the client "fred" issue this command from the master server: bpclient -update -client fred -current_host fred (NetBackup Servers) Description: When browsing files or directories in the "Backup Files" tab of the NetBackup Java client GUI,any file or directory created after September 9, 2001 10:46:SS, will show a date that is in 1973. This behavior will also be seen in the "Task Progress" tab for any backup, archive or restore which was initiated after September 9, 2001 10:46:SS. Although the dates shown on the "Backup Files" tab and the "Task Progress" tab will be incorrect in the cases stated above, it will NOT prevent user backups, archives or restores. This problem does NOT affect the "Restore Files" tab. The dates on the "Restore Files" tab will be the actual date and time the files or directories were last modified and backed up. (NetBackup Servers) Description: When HSM migrated files are restored as migrated files their FHDB entries are not being set active. This can cause data to be lost if migmdclean is run on the HSM volume the data resides on and the volume is recycled. (NetBackup Clients) Description: When backing up a VXFS file system, NetBackup changes the cache state of the files it's backing up in order to facilitate the backup. The cache state will revert to its original value automatically under most VXFS versions. However, with the JFS 3.1 file system, the cache state doesn't automatically revert to its original, so we've modified NetBackup to reset the cache state when the backup of the file is completed. In order to turn on this capability, create the file: "/usr/openv/netbackup/vx_save_cache_state". If this file doesn't exist, NetBackup will execute without manually resetting the cache state (as it did before). (NetBackup clients: solaris2.6) Description: In certain situations usually caused by a disk full condition, the backup image in the NetBackup catalog could become corrupt in such a way as to cause the bpdbm process to crash. In this case, one or more of the following messages would probably be seen in the bpdbm debug log: 12:32:23 [1616.1624] <8> db_get_image_info: no BACKUP_ID 12:32:23 [1616.1624] <8> db_get_image_info: no CREATOR 12:32:23 [1616.1624] <8> db_get_image_info: no SCHED_LABEL 12:32:23 [1616.1624] <8> db_get_image_info: no FILES_FILE Workaround: To workaround the problem, the corrupt backup image should be removed from the NetBackup catalog. (NetBackup Servers) -------------------------------------------------------------------------------- Description: The TLM robotic daemon was not updating its internal tables to reflect the location of a requested volume when it was determined that the requested media was found in another drive that was in use. The associated mount request will now be queued waiting for the volume to become available, instead of being resubmitted immediately, only to fail. (NetBackup Servers) Description: If multiple backups for a file or directory exist and the verbose mode is selected when browsing for restore with the xbp interface, the permissions and timestamps shown for each version of the file or directory will be the same as the most recent backup of that file or directory. The correct data is restored even though the permissions and timestamp display is not correct. Workaround: Use the Java interface, it does not show permissions, but will show the correct modified timestamp. (NetBackup Clients) Description: The bpbrm process intermittently fails to connect to a database client. The bpbrm log will contain an error message similar to the following: 09:07:00 [1892.1936] <16> bpcr_get_socket_rqst4: bpcr_connect timeout during select after 360 seconds followed by: 09:07:00 [1892.1936] <16> bpbrm main: cannot create data socket, A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond. (10060) (NetBackup Servers) Description: Backup jobs that fail right away after being started with an error such as an "exit status code 71: no files found", may be incorrectly assigned an "exit status code 150: termination requested". (NetBackup Servers) Description: To make partial restores easier, it is now possible to exclude items from a restore by adding a '!' character in front of a file or directory path specification. The exclude specifications may appear on the bprestore command or in the bprestore listfile. Once something is excluded, a subset of the files or directories can be re-included by a following include specification. Excludes can also be added when editing the selection list with the bpadm interface. Additional Notes: Restore excludes cannot be used when doing NDMP restores. (NetBackup servers and clients) Description: Attempting to retrieve a known masters list from another host, which is a Global Data Manager (GDM) host, would fail if the local host is not also a GDM host. The functionality was changed to check the GDM status of the other host only, and return the known masters, if configured. This allows non-GDM hosts to examine the known-masters lists of other GDM hosts. (NetBackup Servers) Description: Add support for the CLEAN PERIODIC TapeAlert flag for automated drive cleaning. (NetBackup Media Manager Servers) Description: Add support for IBM 7337, which is a rebadge of an Adic Scalar 100. (NetBackup Media Manager Servers) Description: The bpdbm process and/or the bpsched process would create a core file when attempting to read image file information. This only fails if a bpimage -create_image_list has been done for the client. Workaround: Delete the following files from the db/image/ directories: IMAGE_FILES IMAGE_INFO IMAGE_LIST (NetBackup Servers) Description: The Device Configuration Wizard can now detect and configure the following Breece Hill robotic libraries: Q4.30, Q4.50, and Q6.100. (NetBackup Servers) Description: Jobs that initially queue with a 219 error, due to the ltid process being down on the media server, would remain on the worklist in the done state with an "exit status code 219: the required storage unit is unavailable" error. They would not be removed until the main scheduler process terminated. If the main scheduler process is running 24/7, the job acts like a queued job that never runs. Workaround: Add the following to bp.conf WAIT_IN_QUEUE TIMEOUT_IN_QUEUE = 1 This will cause the job to fail and be removed from the worklist if a 219 exit status code is produced. (NetBackup Servers) Description: Restoring more than 25 AFS volumes at a time, may cause the restore to fail. The progress log will contain a "can't open a pipe: Too many open files" message and TAR will exit with a status of 4. Workaround: Restore fewer AFS volumes at a time. (NetBackup Clients) Description: In rare cases duplication of images may hang. The bptm read and write processes complete, but the bpduplicate process does not terminate. (NetBackup Servers) Description: Autodiscovery of multiple data streams fail in the following cases. - if the client is a DHCP client. The result is that only one stream instead of multiple streams is created for the backup. Workaround: The NEW_STREAM directive may be used to divide the backup into multiple streams. (NetBackup Servers) Description: Under certain circumstances, some robotic processes can erroneously decide there are no drives available for a mount and return an error to the bptm process, who in turn returns error 134 EC_resources_busy to the bpsched process. This was not a fatal error, as the backup would be requeued and run at a later time. For this end case to occur, all drives in the library must be either in use or dismounting (i.e. a tpunmount done on 1 or more drives) and another tape mount request is done before the dismounts are complete. (NetBackup Servers) Description: 3.4 Patch J0850645 introduced a regression in bpbackup processing. An immediate backup, which uses a progress log, must specify a client name via the -h option, rather than using the default. If the client is not specified, the backup aborts. The progress log will contain the following message: bpbackup -i -c 8500_jec -s 8500_sched -t 0 -w -L /tmp/progress EXIT STATUS 25: cannot connect on socket Workaround: For an immediate backup, either run the backup without a progress-log option, or, if the progress-log option is used, then also specify the client option (-h ). (NetBackup Servers) Description: Setting non-blocking I/O on the input side of a socket also causes it to be set on the output side of the socket as well on Solaris 7 and 8, SGI, HP and other non-NT platforms. This results in a EAGAIN error (Resource temporarily unavailable) when trying to write data on the output side of the socket if the system buffers are full. This can happen with TIR backups that use the non-blocking feature on the input side of the socket. A backup status of 24 (socket write failed) or 174 (media manager - system error occurred) will be the final status reported for the backup. Group and user ids for files and directories were not properly handled when numeric values are used instead of names. Numeric values would be replaced by the group and/or user id of the restorer. This is normal behavior for user directed restores but it was happening for root restores as well. Workaround: Turn off the TIR attribute in the backup class. (NetBackup Clients) Description: In some situations a NDMP backup or restore would not handle the NDMP server disconnecting. The symptoms are that the bptm process would "spin" but eventually timeout. (NetBackup Servers) Description: The bptm process could request the files file at the end of the backup before it was completely written. (NetBackup Servers) Description: A path name of greater than 1024 characters causes a backup to abort. (NetBackup Clients) Description: True image restores from a disk storage unit do the restores in a different order which causes the directory timestamps to get set to the restore time instead of the time that they were when the backup was done. (NetBackup servers and clients) Description: Modified license checking for NetBackup Advanced Reporter (NBAR). (NetBackup Servers) Description: Add support for configurable Media ID generation. The LTO barcode format consists of 8 characters nnnnnnLx, where nnnnnn ranges from 0-999999 and x varies from 1-9. At 3.4GA, NetBackup generates Media ID based on the last 6 characters of the media's barcode. Using LTO barcodes will limit the Media IDs generated since the Lx is common for numerous barcodes. A problem can occur when a robotic library contains multiple media with barcodes that have the same last 6 characters. For example, a library has media with barcodes "S00006L1" and "120006L1". The Media ID generated for both pieces of media is identical and when a robot inventory update is performed, the following error occurs: Update failed: could not add new media ID '0006L1' into slot 18 Insert media failed: Media ID not unique to database (34) With NetBackup 3.4 patch NB_34_1 installed on all NetBackup media servers and master server, how Media IDs are generated is now configurable. Media ID generation rules can be added to the vm.conf file to specify the robot number and barcode length, with multiple configuration entries for each robot or for each barcode format. The Media ID generation rule has the following syntax: MEDIA_ID_BARCODE_CHARS ::::: where robot_number is the robot number barcode_length is the number of characters in the media's barcode format c1 is the first character of the Media ID c2 is the second character of the Media ID c3 is the third character of the Media ID c4 is the fourth character of the Media ID c5 is the fifth character of the Media ID c6 is the sixth character of the Media ID c1-c6 can specify a character from the media barcode or specify a fixed character by prefixing the character with a "#". For example, the rule: MEDIA_ID_BARCODE_CHARS 0 8 #N:1:3:4:5:6 generates Media IDs for robot 0 using the character N and the 1st, 3rd, 4th-6th characters of the 8 character barcode, and for the barcode, 123456L1, generates Media ID, N13456. The rule: MEDIA_ID_BARCODE_CHARS 0 8 1:2:3:4:5:6 generates Media IDs for robot 0 using the 1st-6th characters of the 8 character barcode and for the barcode, 006498L2, generates Media ID, 006498. (NetBackup Servers) Description: The scripts: volmgr/tools/vmconf.sh volmgr/tools/vmmedia.sh would exit with hardware type not found or version file corrupt message since the /usr/openv/volmgr/version file contained HP700 or HP800 instead of the HP9000-700 or HP9000-800. (NetBackup Media Manager servers) Description: Add support for a Mac OS X 10.0 NetBackup client. Description: When doing BLIB backups to multiple tape drives, if the workload is not balanced and the media mount timeout is short, when one drive finishes the other one needs to finish before the media mount timeout or the BLIB backups will fail. Workaround: Use a longer media mount timeout. (NetBackup Clients) Description: The Adic Fastor and IBM 3581 with barcode readers will acknowledge a new tape in the library but won't read the barcode until an init_element_status is performed. This will cause frozen tapes if a barcoded tape is inserted and inventoried without initialization. NetBackup will add the tape to the database as an UNKNOWN and assign the tape a mediaid. Then if the robot is powercycled or tldcd performs an init_element_status, NetBackup will think the tape has been swapped out for one with a barcode. Upon mounting the tape, NetBackup will find a label that doesn't correspond to the barcode and will then freeze the tape. Workaround: Use the Adic FastStor or IBM 3581 operator's guide and manually scan barcodes each time a tape is inserted. (NetBackup Servers) Description: If the NetBackup server is setup to disallow client restore requests, the restore request would get denied but a restore job would still get created and appear in the 'queued' state until an administrator kills the job. (NetBackup Servers) Description: Allow NetBackup to cross one mount point, if the file path ends in '/*' and cross mount points is turned off to match NetBackup 3.1.1 behavior. (NetBackup Clients) Description: If a media server is down, non-local drives may be assigned for backups when local drives should be used. (NetBackup Servers) Description: The Device Configuration Wizard can now detect and configure ADIC and IBM SCSI controlled robotic libraries. STK L700 libraries, running firmware which reports the L700 as a SCSI-3 device, are now properly configured. (NetBackup Media Manager Servers) Description: Changes for supporting the ADIC Scalar 10K robot. (NetBackup Servers) Description: Add Device Discovery recognition for entire line of Compaq tape drives. (NetBackup Servers) Description: Add robotic auto cleaning support for STK L series robots (NetBackup Servers) Description: Add recognition for 0x5/0x83/0x10 sense data that has been observed in some tld robotics. (NetBackup Servers) Description: When there are no available drives left for the bpsched process to start more jobs, it would go into a wait state for jobs to finish, freeing up drives to start more. If a drive was unavailable or down when this state is entered, the bpsched process would not recognize that it was available again until a job completed. (NetBackup Servers) Description: Added example Tru64 5.x tape and robot device file names to the tpconfig help text. The vmconf script will now display an error message if it is run on a Tru64 5.x system. (NetBackup Servers) Description: When doing BLIB backups of thousands of files that have not changed, both the client and server processes can get blocked on writes to the socket that go to each other. (NetBackup Clients) Description: If there is more than one storage unit defined for a media server, and the specified hostname for each storage unit represents a different network interface (that is, the media server has more than one network interface card), in a rare circumstance NetBackup may incorrectly reassign tapes. For example, the hostname for storage unit A is specified as "xyz-nic1", which represents one network interface, and the hostname for storage unit B is specified as "xyz-nic2", which represents a different network interface on the same media server. Workaround: On Unix systems, create the file /usr/openv/netbackup/bin/bpsched.d/CHECK_EXPIRED_MEDIA_INTERVAL and enter a 0 (zero) on the first and only line of the file. On an NT server, create the file \NetBackup\bin\bpsched.d\CHECK_EXPIRED_MEDIA_INTERVAL and enter a 0 (zero) on the first and only line of the file. (NetBackup Servers) Description: Group and user ids for files and directories were not properly handled when numeric values are used instead of names. Numeric values would be replaced by the group and/or user id of the restorer. This is normal behavior for user directed restores but it was happening for root restores as well. (NetBackup Clients) Description: Backup '.' files were allowed when file paths of the form .../* are used. This caused mount points at the level .../* to not be crossed if the cross mount points option is off. The method of handling the '.' files was changed so that the first level mount point will always be crossed when specified with a /*. Workaround: Instead of specifying paths using .../* use .../*.* (NetBackup Clients) Description: Add support for the STK SN6000 (Storagenet 6000 Storage Domain Manager). The master server/remote media using the STK SN6000 must have a Multihosted Drives (SSO) license key to use the SN6000. StorageTek (STK) introduced the SN6000 on 10/17/2000 as a method for providing tape drive virtualization. Logical tape drives are presented to standard host operating system interfaces (tape drivers), while robotic control is accomplished through the Automated Cartridge System (ACS). Workaround: Verify the SN6000 has enough physical drives available to handle all mount requests. Use backup windows in class/policy management to avoid over committing the drives. (NetBackup Servers) Description: 1) vopie updates of Mac clients took a much longer time than updates of other client types due to the initial connection to and disconnection from the Mac client. The actual vopie update isn't that time consuming. There isn't much that can be done about the connection delay. However, bpauthsync can be changed to not connect as often to the client. For example, suppose there are four servers and a Mac client to be synced. Previously, bpauthsync would connect to the client four or five times to complete the synchronization. Now, bpauthsync will only connect once to the client to sync with all four servers. 2) bpauthsync could not update the private (unhashed) vopie files, it could only update the public (hashed) vopie files. A new "-always_write" argument has been added to bpauthsync. When "-always_write" is specified with "-vopie", bpauthsync will update both the private and public vopie files. (NetBackup Servers) Description: The bprd process will report "couldn't disconnect from bpcd" errors when disconnecting from bpcd on the client even though bpcd disconnected successfully. An error 26 - "Handshake Failed" message may be seen in a server debug log file. The bprd process is now more robust so that handshake failed errors are less likely to occur when disconnecting from bpcd. (NetBackup Servers) Description: Update SG driver to support auto-sense capability of HBA drivers. By not supporting auto-sense, sense data was being lost when using certain versions of Fibre Channel HBA drivers. This could cause tapes to be marked as bad when they really aren't. System panics have also occurred with certain versions of Fibre Channel HBA drivers. (NetBackup Servers) Description: Adds support for AFS version 3.4 or 3.5 on the AIX 4.3.* platforms. It also includes changes so that if a volume has not had any updates made to it since the last backup, then don't do the "vos backup" and "vos dump" for that volume. This has the potential to drastically reduce the amount of time when backing up AFS volumes. (NetBackup Clients) Description: When doing software compression on an already compressed file that is over 2 gigabytes long, the size information is not set properly during backups which causes the restore to fail with the following message: invalid file size encountered for file /data/abc.Z (NetBackup Clients) Description: Some reports and displays may show an empty timestamp string where one is expected when running under certain locales on certain platforms. For example, running the vmquery command on Linux under the "en_US" locale might display this: # LC_TIME=en_US /usr/openv/volmgr/bin/vmquery -a =========================================================================== media ID: A00000 media type: DLT cartridge tape (11) barcode: -------- description: Added by Media Manager volume pool: NetBackup (1) robot type: TSD - Tape Stacker DLT (10) robot number: 0 robot slot: 4 robot host: hostname volume group: 00_000_TSD created: assigned: last mounted: first mount: expiration date: --- number of mounts: 11 max mounts allowed: --- status: 0x0 =========================================================================== Notice the empty strings for "created:", "assigned:", etc. With this fix, the display will be something like this: # LC_TIME=en_US /usr/openv/volmgr/bin/vmquery -a =========================================================================== media ID: A00000 media type: DLT cartridge tape (11) barcode: -------- description: Added by Media Manager volume pool: NetBackup (1) robot type: TSD - Tape Stacker DLT (10) robot number: 0 robot slot: 4 robot host: hostname volume group: 00_000_TSD created: Fri 10 Nov 2000 03:24:54 PM C assigned: Mon 13 Nov 2000 06:05:47 PM C last mounted: Fri 17 Nov 2000 09:38:17 AM C first mount: Fri 10 Nov 2000 03:31:12 PM C expiration date: --- number of mounts: 11 max mounts allowed: --- status: 0x0 =========================================================================== Workaround: Use the "C" locale if timestamp strings do not appear: # LC_TIME=C /usr/openv/volmgr/bin/vmquery -a (NetBackup Servers) Description: Allow encoding of spaces in filenames. (NetBackup Servers) Description: Backups that need the final server status can fail if the socket that the server status is sent on does not remain active so that the status can be read. (NetBackup Clients) Description: When dealing with a large number of files, a failure can occur with the message "Timed out after X seconds waiting for the file list" in the bpbkar log. The original problem was noted (with VSM) when running mignospace against 10,000 files in a single directory. The original value used for the timeout (FILELIST_TIMEOUT) was 60 seconds. This was increased to 300 seconds. If the file /usr/openv/netbackup/FILELIST_READ_TIMEOUT exists, a value (in seconds) will be read from it. If this value is greater than or equal to 60, it overrides the default. (NetBackup Clients) Description: If the same image is needed by more than 1 restore at the same time, one of the images can fail with: 16:51:07 (9265.007) Status of restore from image created Tue Oct 03 07:31:48 2000 = socket write failed (NetBackup Servers) Description: The ownership of symbolic links restored as root are not changed back to the original uid/gid if the original owner/group no longer exist. Workaround: Put entries in the passwd and group files. (NetBackup Clients) Description: TLH (Tape Library Half-inch) robotic drive addressing may change and become invalid if hardware errors or library reconfiguration occurs while TLH robotic control is active. TLH robotic drive addressing is based on a drive list returned from an IBM Library Device Driver interface. If this drive list changes without being accompanied by a change in the device configuration, mount requests may lead to media being mounted on the wrong drives, or drives may be configured DOWN. Workaround: Run IBM's mtlib utility. Poll the drives every minute and when a drive disappears from the drive list, call tlhcd -t to terminate the NetBackup robotic control daemon. At this point, the administrator will need to reconfigure the devices or get the devices back online before restarting tlhcd. Additional Notes: The TLH robotic test utility, tlhtest, contains three new subcommands: drmapfreeze - Freeze drive address mapping drmapshow - Show drive address mapping drmapclear - Clear drive address mapping In order to take advantage of more a more resilient drive addressing method, the drive mapping must be frozen. Using robtest or tlhtest directly, execute the "drstat" subcommand and verify that the drives being returned in the drive list are consistent with the device configuration. If they are correct, run the "drmapfreeze" tlhtest subcommand. This will "freeze" the mapping between the robot drive number and the IBM device name and number. The current mapping can be displayed with the "drmapshow" subcommand. tlhtest and tlhcd (the TLH robotic control daemon) will then use the frozen drive map (a database) during a mount request to dynamically convert the robot drive number to the correct IBM device number. Once a drive map has been frozen, it remains active until it is cleared with the tlhtest "drmapclear" subcommand, or re-frozen with the tlhtest "drmapfreeze" subcommand. Use the tlhtest "drstat" command, as always, to obtain the robot drive numbers that are used in the device configuration. (NetBackup Servers) Description: Adds support for IBM 3581, 3583, and 3584 tape libraries. Workaround: Setup IBM libraries as TLD robots. Additional Notes: IBM 3583 requires firmware 2.002.0015. (NetBackup Servers) Description: StorEdge Network Foundation HBA requires unique entries in sg.conf and sg.links files. sg.build will detect devices on the HBA and create the correct sg.conf/sg.links entries for them. sgscan will detect devices that are not configured in the sg.conf file and issue a warning message to see the MediaMgr_DeviceConfig_Guide.txt file for configuration details. Additional Notes: Special configuration for the "Sun StorEdge Network Foundation" HBA/Driver The StorEdge Network Foundation HBA requires special configuration to bind device World Wide Port Names for use by the VERITAS SG driver. The script /usr/openv/volmgr/bin/sg.build will add the proper entries to the sg.links and sg.conf files. Before running the sg.build script, make sure that all devices are powered on and connected to the HBA(s). An example of additional entries in /usr/openv/volmgr/bin/driver/sg.conf: name="sg" parent="fp" target=0 lun=0 fc-port-wwn="22000090a50001c8"; name="sg" parent="fp" target=0 lun=1 fc-port-wwn="22000090a50001c8"; An example of the additional entries in /usr/openv/volmgr/bin/driver/sg.links: type=ddi_pseudo;name=sg;addr=w22000090a50001c8,0; sg/c\N0t\A1l0 name="sg" parent="fp" target=0 lun=1 fc-port-wwn="22000090a50001c8"; An example of the additional entries in /usr/openv/volmgr/bin/driver/sg.links: type=ddi_pseudo;name=sg;addr=w22000090a50001c8,0; sg/c\N0t\A1l0 type=ddi_pseudo;name=sg;addr=w22000090a50001c8,1; sg/c\N0t\A1l1 ------------------------------------------------------------------------------ Note: Each time a new device is added or an old device removed, re-create and re-install the new SG configuration (see Configuring the SG/ST Drivers). ------------------------------------------------------------------------------ The script /usr/openv/volmgr/bin/sgscan will check for unconfigured devices, and produce output similar to the following. # #WARNING: detected StorEdge Network Foundation connected devices not in # SG configuration file: # # Device World Wide Port Name 21000090a50001c8 # # See /usr/openv/volmgr/MediaMgr_DeviceConfig_Guide.txt chapter # "Special configuration for "Sun StorEdge Network Foundation" HBA/Driver" # for information on how to use sg.build and sg.install to configure # these devices (NetBackup Servers) Description: The "CRYPT_OPTION = required" configuration entry on a server, will cause a catalog database backup (bpbackupdb) on that server to fail with an error 51. Workaround: Use "CRYPT_OPTION = allowed" instead. (NetBackup Servers) Description: If duplicating SQL images that were not a multiple of 1024 bytes, the write side of bpduplicate would fail. The check for ensuring images are multiples of 1024 bytes has been modified to ensure its a multiple of 512 bytes. Not all of the shared memory used by the bptm process may be released following a duplication. This problem only occurs when using multiplexed duplication. Workaround: Use the ipcrm -m command to free the orphaned shared memory. (NetBackup Servers) Description: Interrupted system call handling was not correct in flush_archive and partial_flush_archive, causing FlashBackup backups to fail when alarm signal was received in the write. (NetBackup Clients) Description: If more than one bpdbm process is writing the class configuration (specifically the /usr/openv/netbackup/db/class//info file), corruption of the configuration may occur. This would result in incorrect values in the attributes section of the class configuration. (NetBackup Servers) Description: Add a filelist directive Microsoft Information Store: \ to support backup of Exchange 2000 (NetBackup Servers) Description: COMPAQ DLT8000 drives are not recognized in auto-configuration. (NetBackup Servers) Description: In a MSCS cluster environment a machine may be known by several names. The tar command may need to know what host name the user originally targeted the restore for (This is currently true for Exchange 2000 in an Active/Active cluster environment). Send to tar via the "-J" option the destination client, requesting client, and browse for source client names. (NetBackup Servers) Description: An SSO installation with a large number of media servers could experience scheduling slow downs when several backups are executed using specific storage units. Queries were made to all the media servers, causing scheduling interruptions that lasted several minutes. Communications are now asynchronous so the scheduling process does not wait for completion. (NetBackup Servers) Description: In certain error conditions, usually right at the end of a backup that was completing, the bptm process could produce a core dump on UNIX or an application exception on NT. (NetBackup Servers) Description: The vmd process (Media Manager volume daemon, NetBackup Volume Manager service) on the DA (device allocation) host core dumps in an SSO environment with more than 15 media servers. If the /volmgr/debug/daemon log file was enabled, the last lines written to the log file would be similar to the following: 09:42:07 [9090] <2> da_register_shared_drives: server read: 5 1 - drive0 drive1 drive2 drive3 drive4 09:42:07 [9090] <2> da_register_shared_drives: returning scan host record for 5 drives registered by host .any.where.com for host Workaround: Limit the number of SSO media servers to 15 or less. Use the MH_SCAN_ABILITY configuration setting and ensure that the scan host is one of the first 15 servers that has registered its drives with vmd/DA, the device allocator. (NetBackup Servers) Description: The unix client, xbp, does not support Japanese localization. (NetBackup Clients) Description: Duplications of True Image Restore (TIR) backups may fail with status 44 (network write failed) when using multiplexed duplication (specified by selecting "Preserve multiplexing when duplicating images" from the Duplicate Images GUI). The duplicate progress log shows messages similar to the following: 13:32:49 INF - Begin writing True Image Recovery information for copy 2. 13:32:51 INF - Duplicate of backupid server_name_0979843140 successful. 13:33:03 INF - Reading True Image Recovery information from media id TLD000. 13:33:03 ERR - Error occurred sending TIR information to socket, No error. 13:33:08 INF - host server_name backup id server_name_0979843139 read failed, network write failed (44). Workaround: The images may be successfully duplicated using non-multiplexed duplication. Do not check the "Preserve multiplexing when duplicating images" on the Duplicate Images GUI. If using the bpduplicate command line, do not specify the -mpx option). (NetBackup Servers) Description: Scheduler fails to requeue the job if a user backup gets an error 134, RESOURCE BUSY. When the job spans a tape, an error 134 can occur while mounting the next tape. Workaround: Touch the file /usr/openv/volmgr/DISABLE_RESOURCES_BUSY (NetBackup Servers) Description: After installing the Encryption product on a master server, the bp.conf file is updated with an "ALLOW_ENCRYPTION" entry. The bpdbm process does not recognize the ALLOW_ENCRYPTION entry. Some NetBackup administration GUIs did not enable the encryption class attribute. Similar problems may also occur with other products that add an "ALLOW_*" entry to bp.conf. Workaround: Stop and restart bpdbm after adding (or deleting) an "ALLOW_*" entry from the bp.conf file or after running an installer that adds (or deletes) an "ALLOW_*" entry from the bp.conf file. (NetBackup Servers) Description: Added support for the following drives in device discovery: Benchmark DLT1, DEC TZ87, DEC TZ88, HP Ultrium, IBM Ultrium. (NetBackup Servers) Description: A file name length of 1023 in a user backup can cause the bpsched process to hang. (NetBackup Servers) Description: If there are a large number of disk storage units, and a large number of disk backups are queued that can run quickly, the scheduler will get tied up processing job complete messages and starting new disk jobs. The request queue is not serviced, causing the sched processes that are building worklists to wait. The result is that tape drives will remain idle, when work should be initiated on them. Workaround: Set the schedule windows of the tape backups to open before the disk backup jobs. They will be loaded on the scheduler work queue before the disk backups that cause the condition. (NetBackup Servers) Description: Enhance support for automatic device discovery/configuration in a clustered environment. (NetBackup Servers) Description: Bpduplicate of TIR backups may fail if the first TIR fragment is on a different media-id than the data, and the mount of the media-id containing the first TIR fragment takes longer than 60 seconds. The error code received is 54 (timed out connecting to client). The bpduplicate progress log shows messages similar to the following: 14:12:43 INF - Waiting for mount of media id CZ1901 on server server_name. 14:14:43 ERR - Unable to create data socket to receive TIR data. 14:15:15 INF - Reading image file information from media id CZ1901. 14:30:47 ERR - Unable to connect to data socket to send TIR data. 14:30:48 INF - host server_name backup id aserver_name_0965758085 read failed, cannot connect on socket (25). 14:30:49 INF - host server_name backupid aserver_name_0965758085 write failed, timed out connecting to client (54). 14:30:51 INF - Duplicate of backup id aserver_name_0965758085 failed, timed out connecting to client (54). (NetBackup Servers) Description: When a tape is broken or a cleaning tape has been misconfigured as a data tape in a TL8 (Tape Library 8mm) robot, NetBackup tries the bad media in all tape drives, downing all of the drives in the robot. (NetBackup Servers) Description: 1) If a user backup is initiated with a progress log, and the master is blocked from client access of the progress log by a firewall, the main bpsched process will hang for client_connect_timout seconds. During this time, no jobs will start. This condition is now blocked by bprd attempting to update the progress log before starting the jobs. The job is failed with an entry in the problems log if the progress log update fails. 2) Extra data streams will be generated if: a class is created with 1 or more schedules, multiple date streams is enabled, a backup is initiated, a schedule type is changed, then another backup is started. The problem was that the streams file ends up with 2 schedules of the same name, with different types. The scheduler would then start all the streams of the same name even though the schedule types were different. 3) The scheduler had a problem handling interrupted system calls correctly when the message queues were full. If a signal occurred while a scheduler process was waiting to write a message on the message queue, the process would exit with an error status. This has been corrected so that the write is attempted again if interrupted system call is detected. 4) The scheduler did not support media servers with older patch levels if the server had multihosted drives. If the scheduler communicated with an older rev media server, it would crash and the reason was difficult to determine. The scheduler will now log an error in the problems log specifying the server of incorrect revision level, and disables multihosted drives for that storage unit. (NetBackup Servers) Description: An import of an image backed up with Multiple Data Streams will fail if imported on another host that does not have Multiple Data Streams enabled for the client. The import receives a "file open failed" error (code 12), but the import does actually import the image which can be successfully used for restores. When this problem occurs, the following message is seen in the import progress log: 09:58:29 INF - Import of class class_name, schedule full (server_name_0967215125) failed, file open failed. The bpdbm log will contain entries similar to the following: 09:58:29 [2078] <16> part_validate: cannot open STREAMS file for client 09:58:29 [2078] <16> validate_image: failure to update STREAMS for client server_name: file open failed (12) 09:58:29 [2078] <4> bpdbm: request complete: exit status 12 file open failed (NetBackup Servers) Description: Provide application error recovery for backups/restores to recover from intermittent fibre channel conditions that have nothing to do with the media itself. Provide accurate KBS/sec message for non-MPX backups when an error recovery occurs. Additional Notes: For application error recovery to be work on HP800, there must be access to the tape drive via a passthru driver. Refer to the NetBackup Media Manager Device Configuration Guide for information how to create device files for passthru for tape drives. (NetBackup Servers) Description: When an NT client is set up to use backup start notify scripts and the execution of the script takes longer than 5 minutes, the backup may fail because the underlying sockets may timeout from under NetBackup. Workaround: Disable multiplexed backups, disable throttling. (NetBackup Servers) Description: Using the bpretlevel command to redefine a retention with a unit of years results in a retention level of infinity. Workaround: Specify months, not years as a unit. Use the interface bpadm to redefine the retention. (NetBackup Servers) Description: The get_license_key utility would display misformatted information for a key's expiration date: Expires: 7 01:00:00, at 2000 instead of Expired: Jul 7, 2000 at 01:00:00 (NetBackup Servers) Description: The bprd process timeouts may occur on client-initiated backups on random clients. The NetBackup server uses a timeout value of 10 seconds for connection requests. In some cases, due to a slow network, the number of hops, or some other reason, the connection may not be established in this time. This sort of failure results in the following messages in the bprd debug log: 07:52:12 [204] <16> db_ERROR: db_begin() failed: cannot connect on socket 07:52:12 [204] <16> logERROR: db_ERROR() failed: cannot connect on socket (25) bprd now looks for the file /usr/openv/netbackup/BPRD_REQUEST_TIMEOUT. If this file exists, bprd uses the value contained in this file as the number of seconds for the connection timeout interval. For instance, if /usr/openv/netbackup/BPRD_REQUEST_TIMEOUT contains the line: 20 then the bprd process times out connection requests after 20 seconds. If /usr/openv/netbackup/BPRD_REQUEST_TIMEOUT does not exist, the bprd process uses a default timeout interval of 300 seconds. This has been increased from the previous default timeout interval, 10 seconds. bprd writes messages to its debug log to display the timeout interval, for instance: 13:22:47 [24038] <4> bprd: Looking for request-timeout file /usr/openv/netbackup/BPRD_REQUEST_TIMEOUT 13:22:47 [24038] <4> bprd: File /usr/openv/netbackup/BPRD_REQUEST_TIMEOUT found, resetting the request-timeout value to 20 seconds If the file /usr/openv/netbackup/BPRD_REQUEST_TIMEOUT does not exist, the debug log contains messages like this: 13:18:03 [23964] <4> bprd: Looking for request-timeout file /usr/openv/netbackup/BPRD_REQUEST_TIMEOUT 13:18:03 [23964] <4> bprd: File /usr/openv/netbackup/BPRD_REQUEST_TIMEOUT not found, using the default request-timeout value of 300 seconds Workaround: Start the backup from the server side. (NetBackup Servers) Description: When running the update_dbclient script, if a particular network package is not installed (primarily on AIX) the 'domainname' command is not found. > ./update_dbclients[195]: domainname: not found. or, a Database client name has a - (dash) in the name. > ./update_dbclients[248]: f2n1-sw_has_conf=yes: not found. > ./update_dbclients[249]: f2n1-sw_conf_modified=yes: not found. > ./update_dbclients[266]: shipped_conf_f2n1-sw=: not found. (NetBackup Servers) Description: 1) When using bprestore to perform an alternate path/alternate client restore from one NT client to another NT client, the alternate path substitution may cause the restore to fail. The error that is displayed in the logs is (WIN32 123: The filename, directory name, or volume label syntax is incorrect. ). This is due to an extra \r character being written to the rename file when it is written on the destination client. 2) When performing an archive operation from a NT client and the NT client is setup to use OTM, the archive log always reports that the OTM license has not been installed on the server regardless of it being installed or not. Workaround: 1) Create the rename file without the original \r character. Most text editors on NT will automatically add this character to the file. Use textpad to allow the rename file to not contain the \r character. 2) There is no workaround to this problem. (NetBackup Servers) Description: Add auto-discovery/configuration support for future SpectraLogic and STK robotic libraries. (NetBackup Servers) Description: The mmcrawl command would core dump when it exceeded a hard limit on the number of host-drive pairs of 256. This is the correct limit for drives per host. But in an SSO configuration, there can be, for example, 50 media servers, each with 20 drives, resulting in 50 * 20 or 1000 drives, which exceeds the limit of 256. Now, all of the hosts and their associated attached drives are detected and the host-drive array size is calculated before malloc'ing space for the array and populated with information. (NetBackup Servers) Description: (1) Bpbkar was core dumping causing exit status 130. (2) The hard link lookup takes too long with a large number of hard links. (3) When doing NFS backups of Network Appliance boxes on Solaris, if async I/O is used on a system with heavy load, the system sometimes returns EINVAL on the aio_read() call. When async I/O is disabled by creating file /usr/openv/netbackup/AIO_READS_MAX with a value of 1 in the file, the sparse file handling code still attempted to do async reads. The sparse file will now do normal reads if the async I/O is disabled. The bpbkar log will have the following message when the problem occurs: 20:28:46 [383] <16> bpbkar dump_sparse: ERR - Read error at byte 233462 reading 1024 bytes in file /data/object/pbrboot.DB. Errno = 22: Invalid argument (4) On HP 11.00 systems, EINVAL was returned on the async_read64 call on the 100th call to async_read64. The logs would show error messages like: 12:57:57 ERR - Read error at byte 6479360 reading 65024 bytes in file /ORACLE/arch/WTCR0022/arch.log1_1123.dbf.Z. Errno = -1: (5) When doing incremental backups on directories with many thousands of files, the default CLIENT_READ_TIMEOUT is not large enough to prevent timeouts when there are very few files that have changed. (6) During large restores, bprd can get an allocation failure when it gets up to 256MB in size on AIX master servers. (7) When doing block incremental restores with archive log images on the same tapes as the incrementals, the restore can hang if it has to wait to restore from the incrementals and starts restoring the archive logs which don't have any dependencies on previous block incremental backups. Workaround: (4) Create file /usr/openv/netbackup/AIO_READS_MAX with a 1 in it on the client. (5) Use a larger CLIENT_READ_TIMEOUT than the default of 5 minutes. (6) Do multiple restores with fewer files in them. (7) Do a separate restore for the archive logs in a separate restore from the block incremental database files. (NetBackup servers and clients) Description: In a multihosted drive configuration with shared drives and non-shared drives in the same robot, jobs can get started to the same drive from multiple storage units, eventually causing media mount timeouts, backup out of window, or a number of other failures. If the condition should occur, bpbrm will return a resource busy status(134) indicating an attempt to use a drive that is already in use on another server. This job will appear in the job monitor as requeued with a 134 status, and the try count will not be incremented. The following lines should be added to bp.conf file if the values of WAIT_IN_QUEUE and TIMEOUT_IN_QUEUE are not already set. WAIT_IN_QUEUE TIMEOUT_IN_QUEUE 300 (NetBackup Servers) Description: Currently, the NetBackup Java GUI (jbp) "logs into" a NetBackup UNIX client (bpjava-msvc) using a UNIX user name and password. bpjava-msvc uses the getpwnam(3) (and possibly getspnam(3)) to authenticate the user name and password. Some UNIX's like Solaris 2.6 and later support Pluggable Authentication Modules (PAM). This makes it possible to support other authentication methods like Kerberos semi-transparently on the system. Applications can be made PAM-aware by calling functions like pam_authenticate(3) and allowing the PAM libraries figure out how to use the configured authentication method. The bpjava-msvc process is not PAM-aware and if the system is configured to use Kerberos authentication, bpjava-msvc will not be able to authenticate a log in attempt because the password value in /etc/shadow is not meaningful. The provided bpjava-msvc will attempt to use PAM authentication first to validate a user name/password. If that fails, it will use the traditional UNIX shadow password. (NetBackup Servers) Description: bpbackup aborts when -w is the last option on the command line. Workaround: Ensure that at least one option follows -w on the bpbackup command line. (NetBackup Clients) Description: Percent complete column is empty in xnbmon list Workaround: Double click on each job to see % complete (NetBackup Servers) Description: Different NDMP server platforms report media errors in different ways. NetBackup did not handle a NDMP_NOTIFY_PAUSED with a reason code of NDMP_MOVER_PAUSE_MEDIA_ERROR. (NetBackup servers ) Description: If a configured NDMP host is not available when NetBackup starts the tldcd and/or tl8cd process will crash in function CloseNdmpFunctions. (Media Manager servers) Description: Bpduplicate fails when duplicating an image from a tape that is in use by another bpduplicate process or is in use by a restore. The bpduplicate fails due to a media open error (error 83). The bpduplicate logs indicate the following error: 14:42:08 INF - Media id B00004 on server server_name is in use, waiting for it. 14:42:30 INF - Beginning duplication on server server_name of client server_name image, creating copy 2. 14:52:12 INF - Waiting for positioning of media id B00004 on server server_name. 14:52:58 INF - host server_name backup id server_name_0961780631 read failed, media open error (83). The bptm logs indicate that another process did a tpunmount of the media-id requested by the bpduplicate process, which causes the open error: 14:52:15 [16782] <2> tpunmount: tpunmount'ing /usr/openv/netbackup/db/media/tpreq/B00004 14:52:15 [16782] <2> getsockconnected: host=server_name.min.ov.com service=bpdbm address=166.98.24.43 protocol=tcp non-reserved port=13721 14:52:15 [16782] <4> mpx_read_backup: successfully restored 1 of 1 requests, read total of 204084 Kbytes at 324.483 Kbytes/sec 14:52:15 [16782] <2> bptm: EXITING with status 0 <---------- 14:52:24 [16873] <2> io_position_for_read: locate block is done 14:52:24 [16873] <2> getsockconnected: host=server_name.min.ov.com service=bpdbm address=166.98.24.43 protocol=tcp non-reserved port=13721 14:52:24 [16873] <16> io_open: cannot open file /usr/openv/netbackup/db/media/tpreq/B00004, No such file or directory 14:52:24 [16873] <2> log_media_error: successfully wrote to error file - 06/ 27/00 14:52:24 B00004 0 OPEN_ERROR Workaround: This problem can be avoided by not having more than one bpduplicate process simultaneously duplicating from the same media-id. Also, do not have a bpduplicate process simultaneously duplicating from a media-id that is being used by a restore. Additional Notes: Added the following I18N messages: SmeBtm2650N "successfully duplicated backup id %s, copy %d, %d kbytes" SmeCom2615N "successfully wrote %d of %d multiplexed duplications, total Kbytes %d at %.3f Kbytes/sec" (NetBackup Servers) Description: The "avrd" process consumes much more CPU time than expected in some SAN media server configurations. Under heavy mount/dismount activity, if there are no locally attached drives configured on a media server, and if the media server is not serving as the scan host for any multihosted drives, the avrd process uses excess CPU time. Workaround: Configure a local drive on SAN media servers for avrd to scan. (NetBackup Servers) Description: Fixed a problem with host name comparison checks that could occur if the hosts being compared had a different number of IP addresses. This would cause problems like a device host would not be added because the host name comparison with an existing device host produced a false match. (NetBackup Servers) Description: If a file is migrated to an HSM nb method volume and the file is renamed the file can not be cached when migstage is used. Workaround: The renamed file may be cached by referencing the file. (NetBackup Clients) Description: Adds smart default media type checking for adding new volumes with inventory robot. (NetBackup Servers) Description: Correct end case problems processing EOM (end-of-media) when writing to NDMP media. In certain cases, such as writing the tar header/trailer during a NDMP backup, if EOM was encountered it would erroneously be interpreted as a media error. (NetBackup Servers) Description: In the bpbrm log on the NetBackup server when either bpbkar or tar is initiated, the following error message will be encountered: ld.so.1: fatal: relocation error: file /opt/lotus/notes/latest/sunspa/libnotes.so: symbol verifyAtomics: referenced symbol not found (NetBackup Clients) README -- Last modified date: Wednesday, October 8, 2003