Re: [Bacula-users] Exchange plugin always truncating logs (when it should not)
On 05/05/12 06:48, James Harper wrote: >> >> I've been using bacula to backup an exchange 2003 server for some time with >> success. You could say I'm an happy user :) >> >> Lately I noticed bacula started truncating logs on EVERY backup, also >> incrementals and differentials, when it should not. >> >> I'm not sure when this happened. I'm inclined to thing it started when >> I upgraded to 5.2.6 But I don't have logs handy till them. I can try to >> reover them from some backup perhaps. >> >> The machine is a Windows 2003 R2 server, 32 bit. I used the installer >> from the project page on sourceforge to install bacula. >> >> If needed I can try installing a previous version on this client and see >> if this makes the problem disappear. Is such a scenario(older client) >> supported? >> > > That's really strange. The code to do this is pretty straightforward... at > job init it does this: > >switch (context->job_level) { >case 'F': > if (context->notrunconfull_option) { > context->truncate_logs = false; > } else { > context->truncate_logs = true; > } > break; >case 'D': > context->truncate_logs = false; > break; >case 'I': > context->truncate_logs = false; > break; >default: > _DebugMessage(100, "Invalid job level %c\n", context->job_level); > return bRC_Error; >} >break; > > Then later it does this: > >if (context->truncate_logs) { > _DebugMessage(100, "Calling HrESEBackupTruncateLogs\n"); > result = HrESEBackupTruncateLogs(hccx); > if (result != 0) { > _JobMessage(M_FATAL, "HrESEBackupTruncateLogs failed with error > 0x%08x - %s\n", result, ESEErrorMessage(result)); > } else { > _JobMessage(M_INFO, "Truncated database logs for Storage Group > %s\n", name); > } >} else { > _JobMessage(M_INFO, "Did NOT truncate database logs for Storage > Group %s\n", name); >} > > And nothing else touches truncate_logs in between that I can see. > > Could you please try doing: > > set debug level=200 trace=1 > > for your fd, run a backup, and then email me the resulting trace file (in the > fd working directory I think). I'm wondering if something has changed and the > callback ordering is different or something. > Thank you for the fast feedback. I'll try this Monday morning and report back. Thank you again! -- Guido Falsi -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Exchange plugin always truncating logs (when it should not)
> > I've been using bacula to backup an exchange 2003 server for some time with > success. You could say I'm an happy user :) > > Lately I noticed bacula started truncating logs on EVERY backup, also > incrementals and differentials, when it should not. > > I'm not sure when this happened. I'm inclined to thing it started when > I upgraded to 5.2.6 But I don't have logs handy till them. I can try to > reover them from some backup perhaps. > > The machine is a Windows 2003 R2 server, 32 bit. I used the installer > from the project page on sourceforge to install bacula. > > If needed I can try installing a previous version on this client and see > if this makes the problem disappear. Is such a scenario(older client) > supported? > That's really strange. The code to do this is pretty straightforward... at job init it does this: switch (context->job_level) { case 'F': if (context->notrunconfull_option) { context->truncate_logs = false; } else { context->truncate_logs = true; } break; case 'D': context->truncate_logs = false; break; case 'I': context->truncate_logs = false; break; default: _DebugMessage(100, "Invalid job level %c\n", context->job_level); return bRC_Error; } break; Then later it does this: if (context->truncate_logs) { _DebugMessage(100, "Calling HrESEBackupTruncateLogs\n"); result = HrESEBackupTruncateLogs(hccx); if (result != 0) { _JobMessage(M_FATAL, "HrESEBackupTruncateLogs failed with error 0x%08x - %s\n", result, ESEErrorMessage(result)); } else { _JobMessage(M_INFO, "Truncated database logs for Storage Group %s\n", name); } } else { _JobMessage(M_INFO, "Did NOT truncate database logs for Storage Group %s\n", name); } And nothing else touches truncate_logs in between that I can see. Could you please try doing: set debug level=200 trace=1 for your fd, run a backup, and then email me the resulting trace file (in the fd working directory I think). I'm wondering if something has changed and the callback ordering is different or something. James -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Exchange plugin always truncating logs (when it should not)
Hi! I've been using bacula to backup an exchange 2003 server for some time with success. You could say I'm an happy user :) Lately I noticed bacula started truncating logs on EVERY backup, also incrementals and differentials, when it should not. I'm not sure when this happened. I'm inclined to thing it started when I upgraded to 5.2.6 But I don't have logs handy till them. I can try to reover them from some backup perhaps. The machine is a Windows 2003 R2 server, 32 bit. I used the installer from the project page on sourceforge to install bacula. If needed I can try installing a previous version on this client and see if this makes the problem disappear. Is such a scenario(older client) supported? The job description is unmodified since february 2011, and I'm sure it was not truncating logs back then. here's the job definition (defaults too): JobDefs { Name = "DefaultJob" Type = Backup Level = Full Accurate = yes Messages = Standard Pool = Default Priority = 10 SpoolData = no Spool Attributes = yes Write Bootstrap = "/var/db/bacula/Bootstraps/%c_%n" Max Start Delay = 20 h Enabled = yes } Job { Name = "nifi10MonthlyMain" Jobdefs = "DefaultJob" Schedule = "MonthlyCycleMain" Storage = "Storageworks-1" Client = "ntfi10-fd" FileSet = "ntfi10 Set" # Exchange plugin not tested with accurate backups. Don't # use the option. Accurate = no Client Run Before Job = "PreBCK.bat before" Client Run After Job = "PreBCK.bat after" } FileSet { Name = "ntfi10 Set" Enable VSS = yes Include { Options { Signature = SHA1 } Plugin = "exchange:/@EXCHANGE/Microsoft Information Store" File = "E:/Exchsrvr/NTFI10.log" File = "E:/Exchsrvr/Old-logs" File = "D:/Save_DB_Marcatempo" File = "D:/Software" File = "D:/tmp_BO" File = "E:/Documenti" File = "E:/backup" } } The accurate and Enable VSS line have always been like that. So I don't think they're problematic. I'm quite sure that up to last year this configuration DID work as expected. Here are the messages from an Incremental: 03-May 23:51 uxifi04-dir JobId 26741: Start Backup JobId 26741, Job=nifi10MonthlyMain.2012-05-03_22.30.01_42 03-May 23:51 uxifi04-dir JobId 26741: Using Device "Ultrium-1" 03-May 23:51 ntfi10-fd JobId 26741: shell command: run ClientRunBeforeJob "PreBCK.bat before" 03-May 23:51 ntfi10-fd JobId 26741: ClientRunBeforeJob: 03-May 23:51 ntfi10-fd JobId 26741: ClientRunBeforeJob: C:\WINDOWS\system32>REM $VGFi: PreBCK.bat,v 1.2 2010/01/19 15:40:24 gfalsi Exp $ 03-May 23:51 ntfi10-fd JobId 26741: ClientRunBeforeJob: 03-May 23:51 ntfi10-fd JobId 26741: ClientRunBeforeJob: C:\WINDOWS\system32>if "before" == "before" goto create 03-May 23:51 ntfi10-fd JobId 26741: ClientRunBeforeJob: 03-May 23:51 ntfi10-fd JobId 26741: ClientRunBeforeJob: C:\WINDOWS\system32>C:\windows\system32\ntbackup.exe backup systemstate /J "NTFI10 SystemState" /F "E:\backup\ntfi10-systemstate.bkf" /R:yes /L:n /SNAP:on 03-May 23:55 ntfi10-fd JobId 26741: ClientRunBeforeJob: 03-May 23:55 ntfi10-fd JobId 26741: ClientRunBeforeJob: C:\WINDOWS\system32>goto :scend 03-May 23:55 ntfi10-fd JobId 26741: ClientRunBeforeJob: 03-May 23:55 ntfi10-fd JobId 26741: ClientRunBeforeJob: C:\WINDOWS\system32>exit 0 03-May 23:55 uxifi04-sd JobId 26741: Spooling data ... 03-May 23:55 ntfi10-fd JobId 26741: Generate VSS snapshots. Driver="Win32 VSS", Drive(s)="ED" 03-May 23:58 ntfi10-fd JobId 26741: Preparing Exchange Backup for NTFI10 03-May 23:58 ntfi10-fd JobId 26741: Truncated database logs for Storage Group First Storage Group 03-May 23:58 ntfi10-fd JobId 26741: Truncated database logs for Storage Group Second Storage Group 03-May 23:58 ntfi10-fd JobId 26741: Truncated database logs for Storage Group Third Storage Group 03-May 23:58 ntfi10-fd JobId 26741: VSS Writer (BackupComplete): "System Writer", State: 0x1 (VSS_WS_STABLE) 03-May 23:58 ntfi10-fd JobId 26741: VSS Writer (BackupComplete): "Microsoft Exchange Writer", State: 0x1 (VSS_WS_STABLE) 03-May 23:58 ntfi10-fd JobId 26741: VSS Writer (BackupComplete): "IIS Metabase Writer", State: 0x1 (VSS_WS_STABLE) 03-May 23:58 ntfi10-fd JobId 26741: VSS Writer (BackupComplete): "MSDEWriter", State: 0x1 (VSS_WS_STABLE) 03-May 23:58 ntfi10-fd JobId 26741: VSS Writer (BackupComplete): "BITS Writer", State: 0x1 (VSS_WS_STABLE) 03-May 23:58 ntfi10-fd JobId 26741: VSS Writer (BackupComplete): "Event Log Writer", State: 0x1 (VSS_WS_STABLE) 03-May 23:58 ntfi10-fd JobId 26741: VSS Writer (BackupComplete): "Registry Writer", State: 0x1 (VSS_WS_STABLE) 03-May 23:58 ntfi10-fd JobId 26741: VSS Writer (BackupComplete): "Removable Storage Manager", Stat