Contents |
| Attention: Sesam doesn't use Archive bit on Windows |
|
Sesam uses now an internal timestamp to decide which files to save during incremental and differential backups |
Sesam disk media pool (DISK_HARD) can now use more than one partition. To enable this function Storage Pools have been introduced. A media pool can be spread over multiple Storage Pools.
Release Notes for version 3.4.1.67
See also the English FAQ or the German FAQ. The use of archive bit may be reset with
See English description 'Backup via VSS' or deutsche Beschreibung 'Sicherung via VSS'.
Sesam can make online backups from Zarafa Groupware Server. In case of restore, single mails or complete user mailboxes can be recovered.
Sesam can save guest computers from VMWare ESX Server.
Sesam can save guest computers from VI3 over VMWare Consolidated Backup
You will find complete CVS change reports here
BUGFIX: MS SQL Server MOVE option: If argument contained a " then part of argument was skipped, e.g. '-a move=testdb:"C:\Inetpub\restore84.mdf"' was shortened to '-a move=C:\Inetpub\restore84.mdf'
BUGFIX: MS SQL Server restore with more then one database file remained in state 'restoring...'
CHANGE: Transfer of big SEL files to RD[S] is now more efficient.
CHANGE: NTBACKUP written media will be accepted
BUGFIX: SBC_WIN32_VERSION=3.2A19: If option '-o vss' was given with non file system backup, e.g. system_state then very large logfiles could be processed and function failed. Now backup is broken due to wrong option.
CHANGE: If source 'ALL' is given and 'nosbc' files is placed under a volume, e.g. 'H:/nosbc' then this volume is skipped from backing up. If volume is explicitely given as target, e.g. 'H:' then 'nosbc' file is ignored --- similar to processing of directories with 'nosbc' file.
BUGFIX: BSR backup with RDS (Remote Device Server) support
BUGFIX: XBSA firewall portrange fixed (since version 3.4.1.41)
BUGFIX: Support for Sesam Windows Server 2000 (again) (Problem was sm_qm_main GetProcessId() ...)
BUGFIX: EFS encrypted files are backed up as normal files since version 3.4.1.61.
BUGFIX: SBCLN Version 3.4.2.4 with correct Transaction Log Processing (was corrupt since version 3.4.2.1).
CHANGE: SBCLN: In case of WINDOWS uppercase of Volume names in paths to allow correct browsing in restore wizard.
CHANGE: sm_recover with extra CLI arguments
CHANGE: SMS Using Global namespace to allow sm_recover access
Licence: Encryption disallowed if ENCRYPT_DISALLOWED 0001 in sm_lic.ini
BUGFIX: Backups on Windows Server 2008 (x64) could fail with message "sm_sms_backup undefined broken"
BUGFIX: GUI User Administration (add or remove) works again.
NEW: Now media with forced tape trailer are accepted for backup. A forced tape trailer appears if a backup stream to the media was aborted by shutdown of the server.
BUGFIX: Installation on Windows Server 2008 (x86) could fail (sm_main).
BUGFIX: Windows: Since version 3.4.1.68 backup throughput could be reduced - resolved with new sbc.exe version 3.2A17.
BUGFIX: If a Windows client was added in GUI then the following error message appeared in Sesam daily protocol: E006-DATABAS[ 768]: CONFIG_CLIENT: empty recordlist, cannot catch requested data.
NEW: sbc.exe 3.2A15 with
NEW: system_state backup without warnings for Active Directory (NTDS) backup on Windows 2003 and Windows 2008
NEW: system_state backup using VSS for Active Directory (NTDS) backup on Windows 2008
NEW: Support for backup with MS Windows Volume Shadow Copy Service.
NEW: GUI backup source browser allows to select 'BSR' if 'BSR' Client software is installed at client side.
NEW: SEPsesam BSR Backup and Desaster Restore for system volume.
NEW: Backup option '-o follow' allows to backup Reparse Points as directories.
NEW: Backup and Restore for backup type VCB_File (vmware)
NEW: Junction dirs and mount points are not longer stored as directories but as reparse points.
BUGFIX: GUI allows to edit restore specific PREPOST files.
BUGFIX: Enhanced maximal path length to 34000, so long path name are handled correctly in GUI restore tree.
CHANGE: (sm_break) Interrupted backups have state '3' now. Backups cancelled before transfering data have state 'c'.
BUGFIX: (sm_config_drives) Don't set number of parallel streams to 5, if standard license is used, but requested streams are less then 5
BUGFIX: Set exit code from 'sm_db export' to >0 if an error occurs. Important during Sesam update to stop it, if export has failed
CHANGE: Preset all non mandatory arguments for 'sm_inventory' and skip all empty arguments during call of 'sm_robot'
BUGFIX: Use function 'int_INSERT_INTO_RESULTS' also for saveset migration, because else 'results.datum' is not set
BUGFIX: (#7609) If original does not exist any more then avoid using unsuccessfully migrated savesets in case of restore.
BUGFIX: Correct check of prio for media_events. Order of 'cal_sheets.object' changed to 'pool drivegrp grpflag drivenum'
BUGFIX: Avoid using same LIS/SGM file name for different backups. Could happen, because of a race condition between select MAX(results.cnt) and INSERT into results. Now 'sm_sepul_event' and 'sm_sbc_com' using same function protected by a semaphore.
CHANGE: SQL Server restore displays SQL Error Messages in Log Level 0.
NEW: sbc.exe displays architecture (x86) or (x64) of program and host
BUGFIX: (sbc.exe) in some error cases during restore output was set to 'Operation successful'
BUGFIX: MS SQLServer database remains in state 'Recovering ...' after restore.
BUGFIX: (sm_pur_status) Wrong entry datum="" in table results could lead to purge of all protocol files.
BUGFIX: During Update PRE, POST, ALARM, DISASTER and NOTIFY interface scripts where fetched from 'downgrade' directory.
NEW: (Windows) If sms.ini [SMS_Server] Compresslevel=1 is set for DISK_HARD media then compressed data is stored till size is reached. Old implementation did not care about compression ratio.
CHANGE: Start subprocess with EXTENDED_STARTUP_INFO for Windows 2008 to avoid 'sm_sms_backup undefined broken' problem.
NEW: LIS file for MS SQL Server with entry for SQL Server datenbase to display it in GUI Restore Wizard.
BUGFIX: Truncate glbv after 512 characters to avoid core dump
BUGFIX: Cut of backup state message after 132 char, before insert it into cal_sheets
BUGFIX: Remove all non ascii characters from CTRLD message, because answer scd-1139: Error: CTRLD reply: Could not find executable..... includes non UTF8 conform bytes, which couldn't be inserted into PostgreSQL DB
BUGFIX: For saveset migration: Copy command could became to long. Now sm_client can handle longer command line.
BUGFIX: If option use empty tape is set in media preferences, check if there is even a tape in drive.
BUGFIX: Check for ALL_JOBS_DONE after all active jobs are set to canceled in database, to be sure that ALL_JOBS_DONE notification is send in all cases
BUGFIX: Selective restore of ESX server backups
BUGFIX: Fix core dump in sm_restore during Open-Xchange selective mail restore.
BUGFIX: Correct name of restore PRE/POST protocol file, so GUI can show it.
BUGFIX: Correct check of higher priority events for media appointments.
BUGFIX: Sesam watch process waits longer before a backup job has to transfer data, because VCB backups need long time for exporting VMDK files.
BUGFIX: Avoid killing non Sesam processes starting with 'sbc' during client shutdown. Attention: Shutdown during a client update will be done by old client !
BUGFIX: Save default ACLs for directories on Linux, which have no additional ACLs.
NEW: Call 'find' with option '-ignore_readdir_race', to avoid errors from 'sbc_find' if file is deleted between 'readdir' and 'stat' calls.
BUGFIX: Correct reconstruct of index files for CYRUS/IMAP public folders.
NEW: Remove OX backup dump files older than 7 days from client.
CHANGE: Faster restore of data from saveset on disk device spreading over more media. This affects migration too.
CHANGE: Don't cut tape device name after first '-' to allow names like '/dev/tape/by-id/LTO-drive' (Linux udev). In former Sesam versions the internal drive number was set by number after '-', but internal drive number is saved in hw_drives.loader_drive.
CHANGE: Adopt read tape trailer to use a tape initialized on Linux on Windows. Especially important for shared drives. Reason is a different number of automatically written file marks written when tape is closed.
BUGFIX: Backups of clients in a task group behind a client, which is locked for backups are started correctly. In Sesam 3.4.1.39 they were silently ignored
BUGFIX: License order is corrected. Because of a missing license check Open-Xchange backup needed a Scalix license.
BUGFIX: Fix core dump during End-of-Media, if more then 10 parallel backups are running.
NEW: The default backup interface can be saved in GUI
BUGFIX: FQDN name couldn't be entered for client name, because filtering invalid characters doesn't accept '.' any more.
NEW: Easy to use template file for sm_disaster interface.
CHANGE: Better log function in SMS. Former log function in Sesam SMS module could cause a slow down in backup speed, especially on Solaris.
CHANGE: Message invalid path specification from sbc for Windows will be interpreted as an error instead of warning.
BUGFIX: Fix core dump during exit of sm_data_server. On Windows sm_data_server sometimes cores during exit log log function.
BUGFIX: Sesam UNIX installation with tar-ball: Because of a change in sm_setup for support of whitespaces in temporary path on Windows, installation on UNIX/Linux failed. Installation by RPM wasn't affected.
BUGFIX: Backups a Windows client from an UNIX Server failed with message SBC logging without final state message, if single port connection was used.
BUGFIX: Avoid deadlock in sm_ctrlc. Sometimes remote access by sm_ctrlc didn't finish properly, causing failed backups.
BUGFIX: Adjust EOL of a medium after saveset copy affected EOL of source pool: The EOL of source medium was set to EOL of target pool. Now EOL of source medium isn't changed any more.
BUGFIX: Media pool could be deleted, even if it was still referenced from a task event: Now reference is checked before.
CHANGE: All database DELETE statements are checked for a valid WHERE clause, to avoid deletion of database records from GUI.
CHANGE: SPARE pools aren't offered as target for backups any more, because they shouldn't be used for regular backups.
BUGFIX: All Sesam clients can be selected as Remote Device Server in Loader Properties: Before only Sesam server could be selected
BUGFIX: Performance data was read to often from a RDS, causing error E001-WATCH Performance data with invalid format. Underlaying problem was ERROR: CTRLD reply: 2008-08-14 19:46:09: scm-1133: Error: Denying service. Maximum active connection reached.
BUGFIX: In GUI->Tasks->Properties: Saving an existing task under a new name and deleting it at once could cause deletion of all tasks not referenced in Sesam database elsewhere.
BUGFIX: New entry in build_host in sm.ini could result in a syntax error during source of sm.ini from shell. This caused failure of some Sesam scripts like sm_os_startup and sm_notify
NEW: OES-Linux can use tsafs to backup eDirectory
CHANGE: Incremental and differential backups of Windows clients using internal timestamp instead of archive bit
NEW: Backup source for Zarafa mail server can be selected by client browser
BUGFIX: A different port for CTRL access can be set in Clients->Properties->Access options again
BUGFIX: Avoid race condition during EOM, which could block all media change operations
BUGFIX: Check stpd listen port for empty or NULL value. Before an empty value wasn't recognized and backup failed because of missing stpd port specification
BUGFIX: Automatic load of tape failed, if no tape was in drive.
BUGFIX: Add package mt-st (including the mt command) as dependency of SEPsesam RHEL RPMs.
BUGFIX: Change '*' and '/' in Zarafa mail subject lines, so GUI can show them correctly
BUGFIX: GUI update function in GUI itself doesn't work
NEW: File system browser creates correct exclude list for NetWare clients
BUGFIX: Create OS group sesam before OS user sesam, because user should become member of group. Only relevant for Sesam server with PostgreSQL as Sesam DB.
CHANGE: write exlude list of SESAM_BACKUP as regular expression for UNIX servers. Avoids that sbc_find walking into work and tmp directory, which can causing errors during SESAM_BACKUP
BUGFIX: GUI Update function works correctly again
BUGFIX: Edit of Sesam interfaces doesn't include trace messages
BUGFIX: Task status screen hangs, if total data amount saved at one Sesam day is larger then 2TB
BUGFIX: Arguments of Command Events are transfered correctly to Sesam client
BUGFIX: Sesam Server on some i386 Linux platforms (RHEL 4, SLES 8...) were linked with shared sqlite library. This causes problems, if no sqlite package was installed on the server too. Now library is linked statically.
Our sales team would be happy to assist you!
Browse and discover videos by SEP.