Patch Name: PHSS_30823 Patch Description: s700_800 10.X OV OB3.50 patch - CS packet Creation Date: 04/07/28 Post Date: 04/08/06 Hardware Platforms - OS Releases: s700: 10.00 10.01 10.10 10.20 s800: 10.00 10.01 10.10 10.20 Products: OmniBackII A.03.50 Filesets: OMNIBACK-II.OMNI-CS,A.03.50 Automatic Reboot?: No Status: General Release Critical: No Path Name: /hp-ux_patches/s700_800/10.X/PHSS_30823 Symptoms: PHSS_30823: Change Request: 8606362207 Backup session cannot be aborted when minimum load balancing equals 2 and first 2 drives report an error. Change Request: H555010860 Scheduled backups do not start. Change Request: H555011989 The post-exec script is started even though the pre-exec script has not been run and the "OB2FORCEPOSTEXEC" option is NOT set. Change Request: H555013036 Pre-exec script does not execute after the Backup Session Manager (BSM) blocked on devices and a post-exec script executes even if session is failing. Change Request: H555013175 When the minimum load balancing is set to 2 and 2 drives are in failure, the session cannot be stopped. Change Request: H555013280 When running the "omnicellinfo -allbdf" command, the following error appears: "Failed to locate disks on host <hostname>! Host may be unreachable." PHSS_29407: Change Request: H555009725 When trying to limit the number of devices used in one session, limitation is not working. Change Request: H555010075 When virtual name does not match the hostname of the host currently running as the cell manager, all the scripts are considered as remote exec scripts. Change Request: H555010103 Imported catalog from media has catalog protection set to None. Change Request: H555010285 Medium import operation fails with message: "|Critical| From: MSM@<host> "datalist" Time:<time> |65:99|Import failed with possible cause: this media already has valid copy in DB." Change Request: H555010298 Backup Session Manager changes Media Agent connection timeout and Media Agent is aborted too fast. Change Request: H555010376 When mount request is issued at the beginning of Informix backup, BackupDeviceIdle timeout expires, Backup session can not be aborted and BSM starts using a lot of memory. Change Request: H555010504 Standard database purge sessions are not started. Change Request: H555010730 crs repeatedly ends abnormally. Change Request: 8606300821 Executing "crs -redebug" from command line displays a usage summary. PHSS_28727: Change Request: H555009297 After aborting host backup the object post exec script does not execute. Change Request: H555009560 Backup run twice when daylight saving time is changed. Change Request: H555009843 While backup with reconnect option is in progress and BSM loses connection to monitor the message appears: "[Warning] From: BSM@<host> "<device>" Time: <time> [61:2013] Some of the backup devices are occupied. Session is waiting for all the devices to get free. [Major] From: BSM@<host> "<device>" Time: <time> [61:3003] Lost connection to Monitor named "noname" on host <host>. Ipc subsystem reports: "IPC Read Error System error: [232] Connection reset by peer" Change Request: H555009863 When backing up very large SAP databases with SSEA split mirror, backups can fail with error: "BR280I Time stamp <time> BR296I Split of mirror disks successful BR274E File /<sapdata_path>/system.data1 not found" PHSS_28021: Change Request: H555008533 Informix backup hangs if a resource type is backed up to a different device than the other resource types. Change Request: H555008690 If a Manager-of-Managers Server is also a MC/ServiceGuard virtual node, then the command: "omnirpt -report list_sessions -timeframe <start_hours_ago><hours_duration> -multicell -no_multiple" lists the matching sessions from this node twice. Change Request: H555008780 Queries from client Cell Server in MoM environment with CMMDB fail with: "[12:1012] Cannot access the file." Change Request: H555008807 When backing up SAP integration with EMC split mirror functionality, not all the SAP files are backed up. Change Request: H555008832 OmniBack internal database backup fails with an error: "[Major] From: BSM@<computer> "<object>" Time: <time> [61:3003] Lost connection to VBDA named <object> on host <host>. Ipc subsystem reports: "IPC Read Error System error: [232] Connection reset by peer" Change Request: H555008927 While configuring OmniBack II reports using GUI, only datalists are listed (barlists are missing). Change Request: H555008892 When performing omnimlist -catalog on media, the program exits and debug.log contains this Interprocess Communication error: 08/06/02 11:13:15 OMNIMLIST.1224.0 ["/src/lib/ipc/ipc.c /main/r31_split/16":3434] A.03.50 b216 [ipc_receiveDataEx] buffer 102400 bytes to small to receive data 258487 bytes => ignored PHSS_26889: Change Request: H555008261 Backup fails with an error: "Internal error. Allocation from library failed. The Media Management daemon reports: Internal error: DbaXXXX functions." Change Request: H555008410 The earlier Patch PHSS_26348 introduced behavior that can cause database integration (for example Oracle or SAP) backups to fail. With PHSS_26348 installed, OmniBack does not properly lock the logical devices to prevent concurrent access by multiple sessions. As a result, database integration backups may fail instead of waiting for the device to become free. Change Request: H555008452 Backup Session Manager does not terminate after the time specified in the SmDaIdleTimeout variable when connection to Disk Agent and Media Agent is lost. Change Request: 8606262839 Configuration report shows an incorrect list of scheduled backups. PHSS_26347: Change Request: H555003134 Device maintenance from command line is not sufficient. Change Request: H555006924 The message for the mount request is not in proper format when forwarded to ITO. Change Request: H555007191 If pre-exec script does not complete in ScriptOutputTimeout time (default 15 minutes), Backup Session Manager lopps and can not be aborted. Change Request: 8606221771 Case sensitive search does not work. Change Request: 8606236143 Not all session manager processes are correctly stopped using csailover.ksh script. Change Request: H555007428 Backup fails with error: "[Major] From: BMA@<HOST> Time:<TIME> [90:64] Can not unload exchanger medium (Storage slot is occupied.)". Change Request: 8606248483 Address of a Cell Console client is incorrectly handled. Change Request: H555007796 omnirpt failes after the installation of PHSS_25002/25003. PHSS_25002: Change Request: B555011718 OmniBack II issues a mount request for tape during the restore session. Change Request: H555005922 Backup is waiting for all pre/post-exec scripts to finish before proceeding. Change Request: H555005975 Purge of filenames aborts daily. Change Request: H555006248 Restore session fails with error: "Can not load exchanger medium.". Change Request: H555006710 Session ends as 'Failed' because OmniBack II tries to back up non existent SAP redo logs. Change Request: H555006771 Load balancing is not working properly. Change Request: H555006799 OmniBack reports that medium copy session finished successfully although it failed. Change Request: H555006810 When OmniBack II Integration backup is aborted, another complete sesson starts and finishes backup of remaining database objects. Change Request: 8606217323 Cell Request Server and session managers frequently abort. PHSS_24658: Change Request: 8606179334 OmniBack II creates duplicated tape labels. Change Request: 8606195237 OmniBack II internal database owner is not properly set after configuring OmniBack II in MC/ServiceGuard environment. Change Request: H555004871 Browsing of a filesystem in Restore GUI sometimes fails with an error: "Cannot browse from root. Please specify starting point.". Change Request: H555005681 Session status is "Completed/Failures" although there is no indication of errors. Change Request: H555005149 If Cell Request Service (CRS) is restarted in cluster aware mode, error: "-2 starting omnidbutil -clear!" occurs in cluster.log file. Change Request: H555005402 Omnitrig fails with an error: "Error parsing schedule purge_schedule.". Change Request: H555005697 Adding OmniBack II licenses in MC/ServiceGuard environment fails with an error: "Cell Server host IP is not in IP range.". Change Request: H555005583 Load balancing uses less devices than expected. Change Request: H555005732 OmniBack II issues mount request, but it can not be confirmed because the tape drive is busy. Change Request: H555005680 Post-exec script specified for a virtual hostname does not start if pre-exec script fails. Change Request: H555005917 Tape import is reported successful although it is not. Change Request: H555005973 Restore session fails with error: "Can not load exchanger medium.". PHSS_23470: Change Request: 8606166806 Session status indicates "Completed" although one of the agents failed. Change Request: B555008272 Devbra reports: "[90:11] Unrecognized option "/" => ignored". Change Request: H555003831 Backup of SAP archive logs fails with: "[Major] From: OB2BAR@<hostname> <ORASID> Time:<time> Can not read BSM message (IPC Read Error System error: [232] Connection reset by peer.)". Change Request: H555004061 Restore does not work correctly after the mount request confirmation. Change Request: H555004021 OmniBack Session Report contains many "[Warning] From: BSM@<hostname> <DBNAME> Time:<time> Received connect from BAR Backup DA on host <host> that is not specified in CLIENT statement in the barlist." warning messages. Change Request: B554000841 Error: "Hostname is not in IP range." is reported after service guard fails over to a secondary host. Change Request: B554000869 Pre- and post-exec scripts return invalid exit status. Change Request: H555004026 After upgrade to OmniBack II 3.50, backup to STK 9840 device fails with an error: "Invalid physical device type => aborting." Change Request: H555004027 Host backup fails if host has more than 600 mountpoints. Change Request: H555004364 If Sybase backup fails before OB2BAR is started, a Major message is reported in the session output, but session status is "Completed". Change Request: H555004452 Objects are lost from OmniBack II internal database after the execution of "omnidbutil -cdbsync" command. Change Request: H555004475 Full Oracle backup can be appended to a medium in a pool with "append incremental only" property. Change Request: - MS Exchange 2000 scheduled backups do not start. Change Request: B554000372 Backup Session Manager is sending wrong SNMP messages to ITO Management server. Change Request: H555005234 When restarting failed sessions multiple sessions are created. Change Request: H555005624 Import of session fails. PHSS_23105: Change Request: H555004350 Bad backup performance after upgrade from OmniBack II 2.55 or 3.00. Change Request: H555004454 Execution of command "omnirpt -report used_media -timeframe <t1> <t2>" results in the following error message: "DB[4] internal error [-16] record not connected to set". "DB[4] internal error [-16] record not connected to set". PHSS_22369: Change Request: 8606144294 Informix barlists are not converted into a OmniBack II 3.50 format after an upgrade from OmniBack II 2.55. Change Request: 8606147687 "List from media" ignores "See private objects" option. Change Request: B554000487 Permanent license shows up as out of IP range. Change Request: B554000496 Backup session is not logged in the database if file pattern search is performed in parallel. Change Request: B554000550 After importing a tape from 2.55 to 3.x cell some of the files can not be restored. Change Request: B554000597 After the upgrade from OmniBack II 3.00 to 3.50, a warning "Received connect from BAR backup DA on host <hostname> that is not specified in CLIENT statement in barlist." appears on a client with multiple lan cards. Change Request: B554000611 A backup of MC/SG can not be aborted after the switch. Change Request: 8606159187 Directories are backed up even if excluded in host backup. Change Request: B554000623 Omnitrig does not pass debugging parameters to omnirpt. Change Request: B554000712 OmniBack II ignores the tape loaded after the mount request for SCSI-II changer device. Change Request: - Option "-change_objname" does not work properly. Change Request: B554000630 List of media fails with an error: "Medium details not available. Medium owned by Cell Manager: <hostname>." Change Request: B554000813 Copy session of write-protected tape is aborted and cleaning tape is requested in the source tape drive. Change Request: B555002886 Internal database server (RDS) abnormally ends due to lack of memory. Change Request: B554000831 Internal database server (RDS) abnormally ends during a SAP restore. Change Request: H555003726 Sessions are disappearing from OmniBack II database. Change Request: H555003740 OmniBack II database gets corrupted. Change Request: H555003729 Sessions are disappearing from OmniBack II database. Change Request: H555003702 Execution of commands "omnidb -session <sess_id>" and "omnirpt -report used_media -timeframe <n> <n>" reports that the database is corrupt, but omnidbcheck reports no problem. Defect Description: PHSS_30823: Change Request: 8606362207 OmniBack II tries to lock new devices. When they are not available, it loops indefinitely. Resolution: If there is an unusable device and minimum load balance is set higher than 1, it is decreased. Change Request: H555010860 Omnitrig process does not start any backup at all if the "holidays" file includes a syntax error. The "holidays" file is read at the beginning, before omnitrig starts the scheduled backups. In case of a syntax error, omnitrig aborts execution, which prevents omnitrig to read the scheduler files and to execute backups at scheduled times. Resolution: Omnitrig behavior in case of syntax errors in the "holidays" file is changed. Now omnitrig continues with an execution in such cases, which means that the syntax errors are skipped, logged to the "debug.log" file and most important scheduled backups are started. Change Request: H555011989 The post-exec script is set to execute if the pre-exec script executes without errors. The problem occurs when the session is aborting and the pre-exec script has not been executed yet as its execution status is still set to OK. Resolution: Added additional checking for session manager state to not execute the post-exec script when the state is set to "BSM_STATE_ABORTED". Part of SSPUX350_091. Change Request: H555013036 Post-exec script may execute even if a pre-exec script has not been executed. The same behavior appears when a session is failing. The pre- and post-exec script behavior differs from documented. Resolution: Pre- and post-exec scripts function as documented. A post-exec script is not executed when session is failing and a pre-exec script has not been executed. Part of SSPUX350_092. Change Request: H555013175 OmniBack II tries to lock devices, but they are not available, so it is looping indefinitely. Resolution: If there is an unusable device and minimum load balance is higher than 1, it is decreased. Change Request: H555013280 "Omnicellinfo" returns random responses. Resolution: "Omnicellinfo" return status changed. PHSS_29407: Change Request: H555009725 Global variable MaxMAperSM is not handled correctly. Resolution: When MaxMAperSM global variable is less then maximum number of devices specified in datalist (load balancing), MaxMAper-SMvalue is used. Change Request: H555010075 When session is configured using post- or pre-exec scripts on virtual machines, no check is being made to determine if IP address is local. Only hostnames (strings) are compared. Resolution: Additional checking is made to determine if an IP address is local. Change Request: H555010103 When importing catalog, its protection is set to None. When catalog is imported with protection None, it is purged out of database first time purge is started. Resolution: Catalog protection is changed to Permanent after importing. Change Request: H555010285 It can happen that media positions remain in database. When you try to import the medium an error occurs. Resolution: New "DbForceImportIfPositionsExist" variable is introduced in global file. When set to 1 it forces medium import even if medium positions exist. Change Request: H555010298 Backup Session Manager changes the timeout value for Media Agent from SmMaIdleTimeout (320 min) to SmPeerId (15) min. This timeout is too low for a BAR session. Resolution: If the session is a BAR session OmniBack uses the timeout value from the SmMaIdleTimeout variable, for other session modes SmPeerId is used. Change Request: H555010376 Omniback II is trying to abort BMA which already has some assigned objects. Resolution: Before aborting BMA Omniback II checks if any objects are assigned to BMA. Part of SSPUX350_090. Change Request: H555010504 Standard purge is not recognized by the scheduler. Resolution: Standard purge is handled in the same way as full database purge. Change Request: H555010730 On HP-UX 11.x systems it is possible that accept() command on a newly created socket fails due to error "ENOBUFS". crs is aborted when such error occurs. Resolution: crs works normally. Change Request: 8606300821 While parsing arguments passed to "crs" binary, missing arguments after "-redebug" options are not expected which causes usage to be displayed. Resolution: When arguments for redebug are not passed on command line, trace file is examined. If trace file does not include parameters, help is written. PHSS_28727: Change Request: H555009297 Before executing the object post-exec script, OmniBack II checks if session is aborting. If it is, then script does not execute. Resolution: Global variable PostExecOnAbortHostBackup was added. If this global variable is set to 1, the post-exec script will execute even if the session is aborting. Change Request: H555009560 The algorithm, that handles the daylight saving time changes is incomplete. Resolution: The algorithm is changed for both time changes: 02h->03h and 03h->02h. Change Request: H555009843 If BSM loses connection to monitor message appears. The problem (reconnect) should only be reported for agents (DA, MA...). Resolution: The function is modified to check if the connection with the agent (DA, MA ...) is broken. If BSM loses connection to monitor, no messages are displayed. Part of SSPUX350_089. Change Request: H555009863 In split mirror environment with very large SAP databases, it can happen that the size of the message buffer that OmniBack II uses is too small to fit in for the complete list of SAP files to be backed up. Resolution: New environment variable is introduced: OB2IPCBUFSIZE. In order to solve this problem, the buffer size can be increased by setting the OB2IPCBUFSIZE environment variable in the .omnirc file on CM, R1 and R2 hosts, specifying the desired amount of memory to be used for the buffer. For determining buffer size please see the Resolution of Id = HSLco37193 below. Part of SSPUX350_088. PHSS_28021: Change Request: H555008533 With Informix integration backup, it is possible to specify which resource type will be backed up by which device. If most of the resource types are backed up to n devices, but the L (Log) resource type is backed up to a separate, additional device, and load balancing is set to (max=n, min=n), then the session will hang because it tries to lock n+1 devices for all the resource type backups. Resolution: OmniBack now unlocks devices, if they were not used for any backup for a certain amount of time. It also stops media agents that have not had any backup agents assigned to them for the timeout period. With both changes (unlock device, stop media agent), it can start new media agents, required by specific resource type backup agents that have already contacted Backup Session Manager. Important note: --------------- In order to get the session running, there are some timeout parameters (global options) that have to be properly set : Global options -------------- A - BackupDeviceIdle B - SmPeerID C - SmDaIdleTimeOut D - SmMaIdleTimeOut Conditions that have to be satisfied : ----------- A<B C<D Example: Default Suggested value value ----------------------------------- A 10min 4min B 5min 5min C 120min 120min D 140min 140min Explanation: ------------ [A<B] After all BAR clients stop sending data to a device, the BSM waits for either A or B timeout to happen for the device. If the timeout A is reached, the backup device will be stopped gracefully, if B timeout is reached, it will be aborted. That's why it is important that the condition A<B is fulfilled. [C<D] This is a general condition that must be met for all MA and DA processes, and is stated in the global options file. It is important to keep the option BackupDeviceIdle under 300 seconds (5 minutes). If 300 or higher, the session behavior will be unpredictable (errors are likely to appear). Part of SSPUX350_087. Change Request: H555008690 When running "omnirpt" command all the cell servers from the "mom_info" and the local Cell Server are queried for matching sessions. The virtual node and the local host are actually the same machine, so the matching sessions are listed twice. Resolution: The function gethostname() is not used any more to determine if the inspected host is also the local host, because in case of a Manager-of-Managers/ServiceGuard environment it returns the physical node hostname and not the virtual node hostname, making the virtual hostname being seen as a non-local host. In this way the virtual hostname is not queried for matching sessions any more. Change Request: H555008780 Format of messages is changed from buffer to pointer but change is not implemented in all areas. Database Session Manager gets message from MoM server in improper format and closes the connection. Resolution: Size of message is dynamically changed, messages are not send in fixed size anymore. Change Request: H555008807 In split mirror environment with huge SAP databases it can happen that the size of the message buffer OmniBack II uses is too small to receive complete list of SAP files to be backed up. Resolution: Setting the OB2IPCBUFSIZE environment variable in the /opt/omni/.omnirc file on Cell Server, R1 and R2 hosts, specifying the desired amount of memory to be used for the buffer solves the described problem. Please read the following instructions and make sure that the patches listed in "Other Dependencies" section (see below) are installed and the environment variable is set on Cell Server, R1 and R2 systems before running the backup! How to determine the buffer size? --------------------------------- N_obj: number of objects (files) N_fs: number of file systems N_lvmvg: number of volume groups N_dg: number of disk groups N_rdsk: number of rawdisks Obj: Object Name ObjBack: Object Backup Name MntPathLen: Path name on backup (R2) host (example: /var/opt/omni/tmp/App_host) fsmnt: File System Mount Point fsdev: File System Device File (example: /dev/dsk/[c0t1d1,lvol1]) vg: Volume group (ex: /dev/vg01) dg: Disk group (VxVM) - dependence on mapfile which is required for creating disk group (example:/dev/vx/dsk/dg01) rdsk: Raw disk file name (example: /dev/rdsk/c1t1d1) AvgLen_<xxx> average length of <xxx> There are two different buffers used: 1. BSM <-> R1 2. R1 <-> R2 Calculating the BSM <-> R1 buffer size -------------------------------------- MsgSize_BSM_R1 = N_obj * (AvgLen_Obj + 1 + MntPathLen) Calculating the R1 <-> R2 buffer size ------------------------------------- MsgSize_R1_R2 = N_obj*(AvgLen_Obj+1+AvgLen_ObjBack) + N_fs*(AvgLen_fsmnt*2 + MntPathLen + AvgLen_fsdev*2)+ N_lvmvg*(AvgLen_vg*2) + N_dg*(AvgLen_dg*2) + N_rdsk*(AvgLen_rdsk*2) Calculating the OB2IPCBUFSIZE value ----------------------------------- OB2IPCBUFSIZE > max (MsgSize_BSM_R1, MsgSize_R1_R2) RESTORE NOTE: ------------- If restore is started using the sapbda utility, the variable OB2IPCBUFSIZE must be exported to environment prior to starting sapbda. Known problems: --------------- If the calculated variable OB2IPCBUFSIZE is correctly set, but is still too small for the file list to fit into the message buffer, the session will fail with the following error seen in the report: BR295I Splitting mirror disks... BR280I Time stamp <time> BR279E Return code from '/opt/omni/lbin/ob2smbsplit /<sap_homedir_path>/.<filename>.lst': 2 BR299E Split/resynchronization of mirror disks failed In this case the variable OB2IPCBUFSIZE must be recalculated to appropriate value. See above how to determine it. Part of SSPUX350_080. Change Request: H555008832 Last message from Disk Agent is sometimes lost. This causes Disk Agent's status reported as "Failed". Resolution: New OmniBack II environment variable: OB2VBDASLEEPBEFORECLOSE is introduced. Its default value is 0. Setting it to a positive number, in /<OMNIBACK_HOME>/.omnirc file, it specifies number of seconds to pass before Disk Agent completes after sending the last block of data. Sleep time insignificantly influences backup performance. Change Request: H555008927 Barlists are members of a group which is not displayed under reporting group in GUI Reporting. Resolution: Barlists are moved to a group that is displayed in GUI. Change Request: H555008892 Media handling results in IPC error when the buffer size is increased and variable OB2IPCBUFSIZE is set as described in above Resolution for problem HSLco37193. Resolution: Media Session Manager is rebuilt in order to properly include the changes introduced with Resolution for HSLco37193. Part of SSPUX350_085. PHSS_26889: Change Request: H555008261 In case of Media Management Database corruption backup can fail due to problems with the media allocation. Resolution: Media allocation is improved. Part of SSPUX350_066. Change Request: H555008410 Two or more integration backup sessions using the same device do not queue for backup, but they all try to use this device at the same time. This results in media agent failure. During integration backup device should be locked for backup, but other sessions can still use it. Resolution: Backup Session Manager correctly locks the device(s), and other sessions wait in the queue for backup until the device is free. Part of SSPUX350_077. Change Request: H555008452 In some cases because of incorrect agent exit code Backup Session Manager does not terminate other agents in correct order, which results in Backup Session Manager hang. Resolution: The exit code is changed. Part of SSPUX350_076. Change Request: 8606262839 Database session manager compares days, not taking into account 15 minute units. Therefore report shows some additional backups, which are scheduled past the specified time frame. Resolution: Beside days Database Session Manager compares 15 minute units. PHSS_26347: Change Request: H555003134 Device enabling, disabling and status report cannot be performed using OmniBack II command line. Resolution: New options "-disable_device", "-enable_device" and "-device" are added to omnimm, omniupload and omnidownload. Part of SSPUX350_038. Change Request: H555006924 The message is not properly filtered out. Resolution: The message is properly filtered before it is send to ITO. Change Request: H555007191 Improper implementation of object assigning to Media Agent causes Backup Session Manager to loop forever. Resolution: Object assignment to Media Agent is corrected. Change Request: 8606221771 Flag for case sensitivity is not checked before search session is executed. Resolution: Case sensitivity flag is always checked. Change Request: 8606236143 The regular expression that is parsed by the awk script does not return the process ID (PID) which should be killed. Resolution: Regular expression string in awk parse command is changed. Change Request: H555007428 Executing an "enter" operation for the slot being used for backup results in backup failure. Resolution: New global option EnterEjectCheckLocking is introduced. If this option is set, Media Management operation checks if slot is locked and does not use it. Part of SSPUX350_072. Change Request: 8606248483 When a Cell Console client connects to the server, user name, system name and domain are parsed correctly, but IP address is not. Resolution: A new environment variable 'OB2CRSSTRICTHOSTNAMECHECKING' is introduced. Setting its value to 1 is to enforce the use of hostnames obtained from IP address. Change Request: H555007796 Improperly built Database Session Manager causes omnirpt to fail. Resolution: Properly built Database Session Manager in included in patch. PHSS_25002: Change Request: B555011718 Rescan before backup option incorrectly deletes locked slots information from OmniBack II database. Resolution: Locked slots are not updated during any rescan. Change Request: H555005922 Backup session is in sleep mode during the pre/post-exec scripts execution. This can cause the session to last a long time. Resolution: New global option is introduced. Details: name: PrePostExecOnEveryVolumeForHostBackup format: PrePostExecOnEveryVolumeForHostBackup=<0 or 1> default: 0 If this option is set (=1), Backup Disk Agent (BDA) will execute pre/post-exec scripts for every volume of host object specified. If this option is not set (=0), Backup Session Manager (BSM) will execute pre-exec script prior to starting first volume BDA and post-exec script after last volume BDA. During the execution of these scripts, the BSM will sleep. Change Request: H555005975 In some special cases it is possible for purge of filenames to fail with error written in the debug.log: "DB[9] internal error [-16] record not connected to set." Resolution: Purge is improved in order to correctly remove such records. Change Request: H555006248 After the mount request during restore from SCSI-II library, OmniBack II tries to load "" medium into drive. Resolution: Restore session manager rechecks the configuration after the mount request is confirmed. Part of SSPUX350_048. Change Request: H555006710 By design, OmniBack II ends session as 'Failed' if there are no objects to back up. Resolution: Two new global variables have been introduced: 1) SessSucessfullWhenNoObjectsBackedUp=0 By setting this variable to 1 user can change the Omniback II behavior in order to end the session as successful when there are no objects backed up. 2) TreatScriptReturnAsNormal=<script return> Setting this variable to a certain number instructs OmniBack II to treat all the script returns equal to the specified number as normal. Part of SSPUX350_062. Change Request: H555006771 There is no warning reported when one creates a datalist or a barlist with load balancing enabled and minimum number of devices greater than number of configured logical devices. Resolution: Comparison of minimum and configured devices number is implemented. Change Request: H555006799 Media Session Manager does not send abort message to Backup Media Agent started after Copy Media Agent failed. Resolution: Media Session Manager properly handles abort message from Copy Media Agent. Change Request: H555006810 1) Error in Backup Session Manager. When OmniBack II - Informix backup aborting, another session starts and finishes the remaining database spaces. 2) Informix wrong error handling. Resolution: 1) Case BSM_STATE_ABORTING was changed in the order that Backup Session Manager aborts session when requested. 2) The way of translating OmniBack II errors into Informix errors was changed with new condition checking. Part of SSPUX350_063. Change Request: 8606217323 Cell Request Server and session managers abort because of an accept() command failure on some HP-UX 11.x systems. Resolution: accept() command is re-executed in case of ENOBUFS failure. Part of SSPUX350_059. PHSS_24658: Change Request: 8606179334 Initializition with "Autolabel" option creates duplicated tape labels. Resolution: Medium initialization does not create duplicated tape labels. Change Request: 8606195237 OmniBack II internal database ownership belongs to primary node instead of floating hostname in MC/ServiceGuard environment. Resolution: Database ownership is set to floating hostname. Change Request: H555004871 The data backed up with an older version of OmniBack II and option "-encrypt" can not be selected from Restore GUI. Resolution: It is possible to select objects backed up with an older version of OmniBack II in Restore GUI. Change Request: H555005681 Message level of a resuming failure in XP environment is set to "Normal" instead of "Major". Resolution: Message level is corrected. Change Request: H555005149 Cell Request Service (CRS) executes "omnidbutil -clear" and returns incorrect exit code. Resolution: Correct exit code is returned. Change Request: H555005402 The parser incorrectly enumerates schedule types. Resolution: Schedule types are properly enumerated. Part of SSPUX350_039. Change Request: H555005697 OmniBack II is not able to list all interfaces in cells with more than 20 LAN cards because the buffer is too small. Resolution: Buffer for storing interfaces is increased. Change Request: H555005583 OmniBack II does not lock all available devices. Resolution: Load balancing locks enough available devices to be used for a session. Change Request: H555005732 When there is no more media available for the running session, mount request is issued. This mount request can not be confirmed because the drive is locked by Backup Session Manager. Resolution: Media Session Manager ignores Backup Session Manager lock making the slot free for mount request confirmation. Change Request: H555005680 OmniBack II improperly recognizes post-exec script as a start request if pre-exec script fails in cluster environment. Resolution: OB2FORCEPOSTEXEC environment variable is added. Setting its value to 1 in the .omnirc file causes post-exec scripts to be started regardless of pre-exec script exit code. Part of SSPUX350_040. Change Request: H555005917 Reporting functionality is missing in Media Session Manager. Resolution: Reporting functionality is added to Media Session Manager. Part of SSPUX350_042. Change Request: H555005973 After the mount request during restore from SCSI-II library, Omniback tries to load "" medium into drive. Resolution: Restore session manager rechecks the configuration after the mount request is confirmed. PHSS_23470: Change Request: 8606166806 If one of the agent fails session status should indicate "Completed/Errors" instead of "Completed". Resolution: Session status is correct. Part of SSPUX350_027. Change Request: B555008272 Database Session Manager starts Device Browser with incorrect options. Resolution: Structure containing Device Browser options is initialized before it is filled with the parameters. Change Request: H555003831 Backup of SAP archive logs fails if a backup request for next archive log comes when Backup Session Manager is already exiting. Resolution: Datalist entries for BAR session are removed. Change Request: H555004061 When restoring the data from STK or GRAU library, wrong medium is loaded into drive after the mount request confirmation. Resolution: Mount request is properly handled. Change Request: H555004021 Backing up Informix Parallel Server with many datafiles where request is coming from the node that is not specified in Informix backup specification results in many warning messages reported during the session. Resolution: New global option InformixparallelBackupIgnoreWarning is added. Warnings are ignored if this option is set to 1. Default value is 0. Change Request: B554000841 Licensing is not working properly for systems with more than 20 lan cards, due to an OmniBack limitation. Resolution: Number of lan cards on a single system can be up to 100. Change Request: B554000869 Pre- and post-exec scripts return invalid exit status because of timing problem. Resolution: Environment variable OB2EXECTIMEOUT is introduced. Variable represents timeout (in seconds) after the end of pre- and post-exec script. Default value is 0. Recommended value is between 5 and 10 seconds. Change Request: H555004026 New Media Class, 9840, is introduced with A.03.50 release. This caused that some of previously created devices did not work with OmniBack II 3.50. Resolution: New utility, mchange, changes media class for a pool and for all devices that use the particular pool. Usage of mchange is: mchange -version | -help mchange -pool PoolName -newtype NewMediaClass Change Request: H555004027 OmniBack II has a 600 mountpoints limitation. Host backup fails if there are more than 600 mountpoints. Resolution: Number of mountpoints per host increased to 50000. Change Request: H555004364 Major messages are not checked by Backup Session Manager. Resolution: Backup Session Manager checks all messages. Change Request: H555004452 If "omnidbutil -cdbsync" command is started in Service Guard environment, some objects disappear from OmniBack II internal database. Resolution: omnidbutil works properly in Service Guard environment. Part of SSPUX350_026. Change Request: H555004475 Backup Session Manager incorrectly appends full Oracle backups. Resolution: Backup Session Manager correctly checks properties of pool. Change Request: - omnitrig does not support MS Exchange 2000 integration. Resolution: Scheduled MS Exchange 2000 backups are properly started. Part of SSPUX350_020. Change Request: B554000372 Normal, Warning and Minor messages can be seen on ITO Management server even though OVfilter is set up to forward only Critical messages. Resolution: Messages are forwarded according to OVfilter setup. Change Request: H555005234 When Backup Session Manager waits for the next client it disables new client connections. Resolution: Proper previously failed session is restarted. Change Request: H555005624 Import of integration session fails if the session is spanned over two or more media. Resolution: Checking for object description is changed. PHSS_23105: Change Request: H555004350 Performance degrated by up to 50% after upgrade from OmniBack II 2.55 or 3.00. Resolution: Redundant update of the objects removed. Part of SSPUX350_023. Change Request: H555004454 Listing of objects failed for the objects that have been purged from the database, because there is no link between objects and object versions. Resolution: Fix is implemented in libob2ecbd.sl. PHSS_22369: Change Request: 8606144294 File /opt/omni/sbin/nb_informix is missing. Resolution: File /opt/omni/sbin/nb_informix is added to OMNI-CS fileset. Change Request: 8606147687 Using "List from media" option in restore context ignores "See private objects" option. Objects that are backed up as "Private" are not listed. Resolution: Option parsing in restore context is improved. Change Request: B554000487 If there are more than 100 network interfaces, ioctl() call fails. Resolution: Buffer for SIOCGIFCONF is increased to support up to 1000 network interfaces. Change Request: B554000496 When backup and file search sessions are started in parallel, file search session locks the database. Backup Session Manager tries to write backup information into locked database. Since this action can not be performed, the following messages are logged in the rds.log: "11740 [10-Apr-00 17:37:07] error status=5 11741 [10-Apr-00 17:37:07] GROUPLOCK ELEMENT: userID=4, LockType=6, LockMode= *, RecSetType=10002, dba=[0:0], odnum=1" Resolution: File pattern search is optimized to minimize the possibilities of group locks. Change Request: B554000550 OmniBack II 3.50 objects are case insensitive which is not true for the SAP objects backed up with 2.55 version. Object tape format is changed with 3.50 as well. These changes caused some SAP objects not to be displayed in the Restore GUI after importing the tape with SAP objects from 2.55 to 3.50. Resolution: Object names from tape are converted into OmniBack II 3.50 format after tape import from 2.55. Change Request: B554000597 System name checking is not performed for systems with multiple LAN cards. Backup Session Manager is checking the system names only when looking for clusters. Resolution: System name checking is performed when looking for clusters and systems with multiple LAN cards. Change Request: B554000611 Aborting backup session is not possible because Backup Session Manager incorrectly aborts Disk Agents after the package switch. Resolution: Disk Agents are properly aborted. Change Request: 8606159187 In host backup only mount point directories can be excluded. If a directory is not a mount point then it cannot be excluded from the backup. Resolution: Non-mountpoint directories can now be excluded. Change Request: B554000623 Scheduled reports are not started with "-debug" option although omnitrig is. Resolution: Command line option passing is implemented for omnirpt. Change Request: B554000712 Mount request is not properly implemented for SCSI-II changer devices. Resolution: OmniBack II properly handles mount requests. Change Request: - Option "-change_objname" is not meant to be a part of OmniBack II A.03.50. Resolution: Option "-change_objname" is removed. Change Request: B554000630 After upgrading OmniBack ServiceGuard package from 3.00 to 3.50, it is not possible to list media due to incorrect ownership of medium. Resolution: Virtual hostname is used for medium ownership. Part of SSPUX350_001. Change Request: B554000813 Copy session is aborted when copying a write-protected tape. When write-protected tape in the source drive is checked, the cleaning tape for that drive is requested and the following message appears: "Cleaning tape requested by device <device_name>. Please clean the drive.". Resolution: The cleanme signal is now processed correctly. Part of SSPUX350_004. Change Request: B555002886 Execution of command "omnidb -filesystem <FS_Object> -file <File_Name>" causes a memory leak in RDS (approx. 300 bytes). If the command is started more than 100000 times, it can cause RDS to die due to a lack of memory. Resolution: Fix is implemented in libob2ecdb.sl. Part of SSPUX350_006. Change Request: B554000831 RDS dies during a large SAP restore due to memory leak described in NSMex02584. Resolution: Fix is implemented in libob2ecbd.sl. Part of SSPUX350_006. Change Request: H555003726 Media sync does not work properly when table of exported/recycled media gets too big (>3000 media). Some new sessions dissapear from database. No error is reported. Resolution: Improved error handling logs this event in debug.log. Part of SSPUX350_003. Change Request: H555003740 Database stays in the hot mode after issuing an "omnidbcheck" command. If a large amount of data is written before OmniBack II daemons are stopped, it can happen that hot files are not properly flushed. This causes corruption of database. Resolution: omnidbcheck flushes hot files immediately after database check. Part of SSPUX350_003. Change Request: H555003729 Media Session Manager is incorrectly initializing media in ServiceGuard environment. Owner of medium is set to hostname of the active node. This results in wrong medium handling and objects from such medium are removed from the OmniBack II database each time Media Session Manager is started. Resolution: Initialization of media in ServiceGuard environment is improved. Before applying the fix for NSMex03737 please run following commands: omnidbutil -change_cell_name <hostname1> and omnidbutil -change_cell_name <hostname2> where <hostname1> and <hostname2> are names of ServiceGuard nodes. Part of SSPUX350_010. Change Request: H555003702 OmniBack II reports an error: "Database is probably corrupt." for backup sessions that span over two or more media if one of them is already overwritten/exported because then some of the objects are already removed from the database. Resolution: Removed objects are skipped during the database queries. Part of SSPUX350_010. SR: 8606362207 H555013175 H555011989 H555013036 H555013280 H555010860 H555010504 H555010730 8606300821 H555010075 H555010285 H555010376 H555010298 H555010103 H555009725 H555009863 H555009297 H555009560 H555009843 H555008927 H555008892 B554000611 B554000630 B554000712 B554000622 B554000550 B554000650 B554000623 B554000830 B554000831 B554000813 B554000418 B554000751 B554000761 B554000556 B554000487 B554000496 B554000776 B554000867 B554000597 B554000887 B554000888 B554000889 B555002886 H555008533 H555008832 H555004477 H555008807 H555008780 H555008690 H555008261 8606262839 H555008452 H555008410 B554000768 8606221771 8606248483 H555007191 H555003134 H555007796 H555007428 H555006924 8606236143 H555005975 H555005922 H555006248 B555011718 H555006771 H555006710 H555006799 H555006810 8606217323 H555005697 H555005402 H555005732 H555005680 H555005149 H555004871 H555005681 H555005583 H555005917 H555005973 8606179334 8606195237 H555004061 H555004026 H555004027 B555008272 H555005234 H555004452 H555003831 B554000869 H555005624 H555004364 H555004475 B554000372 B554000841 H555004021 8606166806 H555004350 8606159187 8606144294 8606147687 Patch Files: /opt/omni/lbin/crs /opt/omni/lbin/bsm /opt/omni/lbin/dbsm /opt/omni/lbin/msm /opt/omni/lbin/mmd /opt/omni/lbin/rsm /opt/omni/lbin/psm /opt/omni/lbin/dbbda /opt/omni/sbin/omnitrig /opt/omni/sbin/omnidbutil /opt/omni/sbin/omnidbcheck /opt/omni/sbin/nb_informix /opt/omni/sbin/utilns/mchange /opt/omni/sbin/install/omniforsg.ksh /opt/omni/lib/libob2ecmn.sl /opt/omni/lib/libob2ecdb.sl /opt/omni/lib/libob2emmdb.sl /opt/omni/lib/libob2eadm.sl /opt/omni/newconfig/etc/opt/omni/options/global /opt/omni/newconfig/etc/opt/omni/sg/csfailover.ksh what(1) Output: /opt/omni/lbin/crs: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:24 :05 2004 /opt/omni/lbin/bsm: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:20 :16 2004 /opt/omni/lbin/dbsm: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:28 :54 2004 /opt/omni/lbin/msm: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:23 :12 2004 /opt/omni/lbin/mmd: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:25 :46 2004 /opt/omni/lbin/rsm: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:21 :53 2004 /opt/omni/lbin/psm: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:29 :29 2004 /opt/omni/lbin/dbbda: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:33 :17 2004 /opt/omni/sbin/omnitrig: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:33 :44 2004 /opt/omni/sbin/omnidbutil: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:27 :58 2004 /opt/omni/sbin/omnidbcheck: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:28 :09 2004 /opt/omni/sbin/nb_informix: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:36 :17 2004 /opt/omni/sbin/utilns/mchange: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:34 :09 2004 /opt/omni/sbin/install/omniforsg.ksh: None /opt/omni/lib/libob2ecmn.sl: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:15 :04 2004 /opt/omni/lib/libob2ecdb.sl: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:16 :56 2004 /opt/omni/lib/libob2emmdb.sl: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:25 :22 2004 /opt/omni/lib/libob2eadm.sl: HP OpenView OmniBack II A.03.50; internal build PHSS _30823/PHSS_30824, built on Tue Jul 20 17:27 :11 2004 /opt/omni/newconfig/etc/opt/omni/options/global: None /opt/omni/newconfig/etc/opt/omni/sg/csfailover.ksh: None cksum(1) Output: 4259444021 735108 /opt/omni/lbin/crs 1625694231 1182707 /opt/omni/lbin/bsm 248477049 1247769 /opt/omni/lbin/dbsm 706797106 694260 /opt/omni/lbin/msm 3988518876 530043 /opt/omni/lbin/mmd 685423602 874939 /opt/omni/lbin/rsm 347185803 435578 /opt/omni/lbin/psm 909021093 365783 /opt/omni/lbin/dbbda 3505324970 415000 /opt/omni/sbin/omnitrig 1554182368 702067 /opt/omni/sbin/omnidbutil 683931110 472364 /opt/omni/sbin/omnidbcheck 4167109757 332670 /opt/omni/sbin/nb_informix 532802823 213672 /opt/omni/sbin/utilns/mchange 3275622251 5084 /opt/omni/sbin/install/omniforsg.ksh 648523381 812458 /opt/omni/lib/libob2ecmn.sl 2239015052 889556 /opt/omni/lib/libob2ecdb.sl 4039632763 651912 /opt/omni/lib/libob2emmdb.sl 1699542391 692720 /opt/omni/lib/libob2eadm.sl 2258981330 34415 /opt/omni/newconfig/etc/opt/omni/options/ global 1805127806 2350 /opt/omni/newconfig/etc/opt/omni/sg/ csfailover.ksh Patch Conflicts: None Patch Dependencies: None Hardware Dependencies: None Other Dependencies: Some of the defect fixes contained in this patch depend on the following other patches or their successors in order to completely resolve the problem: PHSS_28725: Cell Console (CC) packet patch, release Date: Available Change Request: H555009863 Change Request: H555003134 Change Request: H555008807 PHSS_28725 is required for these defect fixes. PHSS_28578: SureStore Agent (SSEA) packet patch, release Date: Available Change Request: H555009863 PHSS_28578 is required for this defect fix. PHSS_28570: Core packet patch, release Date: Available Change Request: H555009863 Change Request: H555008807 Change Request: H555003134 Change Request: B554000768 Bad catalog access messages are displayed without PHSS_28570. PHSS_28019: EMC Agent packet patch, release Date: Available Change Request: H555008807 PHSS_28019 is required for this defect fix. PHSS_28017: SAP packet patch, release Date: Available Change Request: H555009863 Change Request: H555006710 Change Request: H555008807 PHSS_28017 is required for these defect fixes. PHSS_25950: Informix packet patch, release date: Available Change Request: H555006810 PHSS_25950 is required for this defect fix. Supersedes: PHSS_22369 PHSS_23105 PHSS_23470 PHSS_24658 PHSS_25002 PHSS_26347 PHSS_26889 PHSS_28021 PHSS_28727 PHSS_29407 Equivalent Patches: PHSS_30824: s700: 11.00 11.11 s800: 11.00 11.11 Patch Package Size: 3180 KBytes Installation Instructions: Please review all instructions and the Hewlett-Packard SupportLine User Guide or your Hewlett-Packard support terms and conditions for precautions, scope of license, restrictions, and, limitation of liability and warranties, before installing this patch. ------------------------------------------------------------ 1. Back up your system before installing a patch. 2. Login as root. 3. Copy the patch to the /tmp directory. 4. Move to the /tmp directory and unshar the patch: cd /tmp sh PHSS_30823 5a. For a standalone system, run swinstall to install the patch: swinstall -x autoreboot=true -x match_target=true \ -s /tmp/PHSS_30823.depot By default swinstall will archive the original software in /var/adm/sw/patch/PHSS_30823. If you do not wish to retain a copy of the original software, you can create an empty file named /var/adm/sw/patch/PATCH_NOSAVE. WARNING: If this file exists when a patch is installed, the patch cannot be deinstalled. Please be careful when using this feature. It is recommended that you move the PHSS_30823.text file to /var/adm/sw/patch for future reference. To put this patch on a magnetic tape and install from the tape drive, use the command: dd if=/tmp/PHSS_30823.depot of=/dev/rmt/0m bs=2k Special Installation Instructions: PHSS_30823: NOTE! This Patch has to be installed on the Cell Manager host. OmniBack II Services must be shut down before the patch is installed. Use the following command to shut down the services: "/opt/omni/sbin/omnisv.sh stop". After the installation is complete, restart OmniBack II Services using "/opt/omni/sbin/omnisv.sh start". Patch includes the following Site Specific Patches: SSPUX350_001 SSPUX350_003 SSPUX350_004 SSPUX350_006 SSPUX350_010 SSPUX350_020 SSPUX350_023 SSPUX350_026 SSPUX350_027 SSPUX350_038 SSPUX350_039 SSPUX350_040 SSPUX350_042 SSPUX350_048 SSPUX350_059 SSPUX350_062 SSPUX350_063 SSPUX350_066 SSPUX350_072 SSPUX350_076 SSPUX350_077 SSPUX350_080 SSPUX350_085 SSPUX350_087 SSPUX350_088 SSPUX350_089 SSPUX350_090 SSPUX350_091 SSPUX350_092 A Site Specific Patch is a temporary hotfix that you might have received from HP Support if you have encountered a problem with Data Protector for which no General Release patch was or is available yet. If any newer Site Specific Patch - that is one with a higher number - for the Cell Manager is installed on your system, then you must re-apply this SSP after installing this Cell Manager patch.