Re: [Bacula-users] please help with understanding WildDir
On Wed, 3 Jun 2009 07:59:00 -0500, Mike Eggleston said: Morning, I have been running bacula for quite some time (wonderful system). I'm tweaking my backups to reduce just how much stuff is backed up and eliminate those things that can be reinstalled rather than restored if needed. Today I'm looking at a series of files that are copied from an internal source code repositiory. Since the source code repository is backed up nightly, I don't want to backup up an individual's copies of those files (the individuals are by company policy required to manage any files checked out that have changes that are not yet checked back in to the source code repository). In my bacula-dir.conf I have a line: WildDir = /opt/data/*/p4 In last night's backup I have the following file that was backed up: (query, 12, $jobid) /opt/data/user/p4/user_shared/product/dist_sdk/PluginSDK/src/plugins/CaeXML/x64/Debug/runtimeWrite.obj My understanding, that I must have wrong, is that since the bacula-dir.conf this section of the File Set has Exclude = yes that /opt/data/*/p4 is excluded that also implies that anything under /opt/data/*/p4 is also excluded. I do have a statement further down in the File Set section of 'File = /opt/data' so that everything else is included. Is there a conflict in what I have done? Is there a better way to exclude part and include the rest? Your approach should work, but the description of your fileset was too vague for me to tell why it doesn't. I recommend that you post the complete fileset definition to the list. __Martin -- OpenSolaris 2009.06 is a cutting edge operating system for enterprises looking to deploy the next generation of Solaris that includes the latest innovations from Sun and the OpenSource community. Download a copy and enjoy capabilities such as Networking, Storage and Virtualization. Go to: http://p.sf.net/sfu/opensolaris-get ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] please help with understanding WildDir
On Wed, 3 Jun 2009 14:20:43 -0500, Mike Eggleston said: On Wed, 03 Jun 2009, Martin Simmons might have said: On Wed, 3 Jun 2009 07:59:00 -0500, Mike Eggleston said: Morning, I have been running bacula for quite some time (wonderful system). I'm tweaking my backups to reduce just how much stuff is backed up and eliminate those things that can be reinstalled rather than restored if needed. Today I'm looking at a series of files that are copied from an internal source code repositiory. Since the source code repository is backed up nightly, I don't want to backup up an individual's copies of those files (the individuals are by company policy required to manage any files checked out that have changes that are not yet checked back in to the source code repository). In my bacula-dir.conf I have a line: WildDir = /opt/data/*/p4 In last night's backup I have the following file that was backed up: (query, 12, $jobid) /opt/data/user/p4/user_shared/product/dist_sdk/PluginSDK/src/plugins/CaeXML/x64/Debug/runtimeWrite.obj My understanding, that I must have wrong, is that since the bacula-dir.conf this section of the File Set has Exclude = yes that /opt/data/*/p4 is excluded that also implies that anything under /opt/data/*/p4 is also excluded. I do have a statement further down in the File Set section of 'File = /opt/data' so that everything else is included. Is there a conflict in what I have done? Is there a better way to exclude part and include the rest? Your approach should work, but the description of your fileset was too vague for me to tell why it doesn't. I recommend that you post the complete fileset definition to the list. __Martin Thanks, Martin. Here's the file set. I didn't want to send a huge message to the list. Looking at the File Set I see I misspelled backula. Great! That one change will reduce the size of the backup on the bacula host. Mike File Set { Name = Unix-Files Ignore Fileset Changes = yes Include { It should work if you insert Options { Exclude = yes WildDir = /opt/data/*/p4 } here. Options { signature = MD5 } File = / File = /boot File = /home File = /opt File = /opt/CVS File = /opt/backups File = /opt/bacula File = /opt/pointwise File = /opt/pwi File = /etc File = /var File = /usr/share File = /var/lib File = /var/lib/imap File = /var/lib/ldap File = /var/lib/mysql File = /var/spool File = /var/spool/imap File = /var/spool/mail File = /var/spool/samba File = /var/www File = /var/www/html File = /var/www/html/bugzilla File = /var/yp Are these all separately mounted filesystems? If not, you'll get duplication when Bacula walks down from /opt into /opt/bacula etc. __Martin -- OpenSolaris 2009.06 is a cutting edge operating system for enterprises looking to deploy the next generation of Solaris that includes the latest innovations from Sun and the OpenSource community. Download a copy and enjoy capabilities such as Networking, Storage and Virtualization. Go to: http://p.sf.net/sfu/opensolaris-get ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] compiling bacula v3.0.1 on FreeBSD v7.1
On Mon, 8 Jun 2009 15:51:07 -0400, Lars Nordin said: Anyone out there tried compiling bacula v.3.0.1 on FreeBSD v7.1? When I try I get this error: Compiling jcr.c jcr.c: In function 'void _dbg_print_jcr(FILE*)': jcr.c:1067: error: cast from 'pthread*' to 'int' loses precision *** Error code 1 Is this 64-bit FreeBSD? It looks like a bug to me, because the code is casting pthread_t to int. It probably doesn't complain on most 32-bit systems, but it isn't allowed by the standard. As a hackaround, trying changing (int) to (int)(long) on line 1067 of src/lib/jcr.c. That is not correct either, but it might remove the error. I suggest you file a bug report. __Martin -- Crystal Reports - New Free Runtime and 30 Day Trial Check out the new simplified licensing option that enables unlimited royalty-free distribution of the report engine for externally facing server and web deployment. http://p.sf.net/sfu/businessobjects ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] to-disk differentials now using multiple volumes?
On Thu, 11 Jun 2009 10:43:43 -0400, Jeff Dickens said: Bumping my own note here.. Does anyone know why things might be different under 3.0 with this end of media thing? The message just before was: End of Volume ted2-diff-0003 at 0:2111477919 on device ted2-filedev (/bacula/ted2). Write of 64512 bytes got 1888. This suggents your disk is (or was) full. Check the syslogs on the SD machine. __Martin Jeff Dickens wrote: I'm using pretty much the same configuration for my differential backups as I did under Bacula 2.2.8, but for some reason under 3.0.1 it's getting end of media on a disk volume and moving to a new volume. There seems to be some kind of new behavior here.. It's possible the storage disk could have run out of space. Could anyone explain what is happening here? 05-Jun 22:34 packrat-dir JobId 31991: Start Backup JobId 31991, Job=ted2.2009-06-05_22.20.01_15 05-Jun 22:34 packrat-dir JobId 31991: Purging oldest volume ted2-diff-0003 05-Jun 22:34 packrat-dir JobId 31991: 1 File on Volume ted2-diff-0003 purged from catalog. 05-Jun 22:34 packrat-dir JobId 31991: There are no more Jobs associated with Volume ted2-diff-0003. Marking it purged. 05-Jun 22:34 packrat-dir JobId 31991: All records pruned from Volume ted2-diff-0003; marking it Purged 05-Jun 22:34 packrat-dir JobId 31991: Using Device ted2-filedev 05-Jun 22:34 ted2-fd JobId 31991: DIR and FD clocks differ by -39 seconds, FD automatically compensating. 05-Jun 22:34 packrat-sd JobId 31991: Recycled volume ted2-diff-0003 on device ted2-filedev (/bacula/ted2), all previous data lost. 05-Jun 22:34 packrat-dir JobId 31991: Volume used once. Marking Volume ted2-diff-0003 as Used. 05-Jun 22:34 ted2-fd JobId 31991: Generate VSS snapshots. Driver=VSS WinXP, Drive(s)=C 05-Jun 22:40 packrat-sd JobId 31991: End of Volume ted2-diff-0003 at 0:2111477919 on device ted2-filedev (/bacula/ted2). Write of 64512 bytes got 1888. 05-Jun 22:40 packrat-sd JobId 31991: End of medium on Volume ted2-diff-0003 Bytes=2,111,477,920 Blocks=32,730 at 05-Jun-2009 22:40. 05-Jun 22:40 packrat-dir JobId 31991: There are no more Jobs associated with Volume ted2-diff-0448. Marking it purged. 05-Jun 22:40 packrat-dir JobId 31991: All records pruned from Volume ted2-diff-0448; marking it Purged 05-Jun 22:40 packrat-dir JobId 31991: Recycled volume ted2-diff-0448 05-Jun 22:40 packrat-sd JobId 31991: Recycled volume ted2-diff-0448 on device ted2-filedev (/bacula/ted2), all previous data lost. 05-Jun 22:40 packrat-dir JobId 31991: Volume used once. Marking Volume ted2-diff-0448 as Used. 05-Jun 22:40 packrat-sd JobId 31991: New volume ted2-diff-0448 mounted on device ted2-filedev (/bacula/ted2) at 05-Jun-2009 22:40. 05-Jun 22:48 packrat-sd JobId 31991: End of Volume ted2-diff-0448 at 1:152619127 on device ted2-filedev (/bacula/ted2). Write of 64512 bytes got 1928. 05-Jun 22:48 packrat-sd JobId 31991: End of medium on Volume ted2-diff-0448 Bytes=4,447,586,424 Blocks=68,942 at 05-Jun-2009 22:48. 05-Jun 22:48 packrat-dir JobId 31991: There are no more Jobs associated with Volume ted2-diff-0449. Marking it purged. 05-Jun 22:48 packrat-dir JobId 31991: All records pruned from Volume ted2-diff-0449; marking it Purged 05-Jun 22:48 packrat-dir JobId 31991: Recycled volume ted2-diff-0449 05-Jun 22:49 packrat-sd JobId 31991: Recycled volume ted2-diff-0449 on device ted2-filedev (/bacula/ted2), all previous data lost. 05-Jun 22:49 packrat-dir JobId 31991: Volume used once. Marking Volume ted2-diff-0449 as Used. 05-Jun 22:49 packrat-sd JobId 31991: New volume ted2-diff-0449 mounted on device ted2-filedev (/bacula/ted2) at 05-Jun-2009 22:49. 05-Jun 22:52 ted2-fd JobId 31991: VSS Writer (BackupComplete): MSDEWriter, State: 0x1 (VSS_WS_STABLE) 05-Jun 22:52 ted2-fd JobId 31991: VSS Writer (BackupComplete): Microsoft Writer (Bootable State), State: 0x1 (VSS_WS_STABLE) 05-Jun 22:52 ted2-fd JobId 31991: VSS Writer (BackupComplete): Microsoft Writer (Service State), State: 0x1 (VSS_WS_STABLE) 05-Jun 22:52 ted2-fd JobId 31991: VSS Writer (BackupComplete): WMI Writer, State: 0x1 (VSS_WS_STABLE) 05-Jun 22:53 packrat-sd JobId 31991: Job write elapsed time = 00:18:24, Transfer rate = 7.030 M bytes/second 05-Jun 22:53 packrat-dir JobId 31991: Bacula packrat-dir 3.0.1 (30Apr09): 05-Jun-2009 22:53:26 Build OS: i686-redhat-linux-gnu redhat JobId: 31991 Job:ted2.2009-06-05_22.20.01_15 Backup Level: Differential, since=2009-05-21 04:00:04 Client: ted2-fd 2.2.6 (10Nov07) Linux,Cross-compile,Win32 FileSet:XP_WS_OL 2009-05-17 17:18:47 Pool: ted2-pool-diff (From Job DiffPool override) Catalog:MyCatalog (From Client resource) Storage:ted2-filestore (From Job
Re: [Bacula-users] Performance with MySQL queries since 3.0.0 (Dir inserting attributes hang)
On Tue, 16 Jun 2009 15:05:18 +0200, Tom Sommer said: Hi, I have a somewhat pressing problem with the performance of my Bacula installation. My MySQL database currently holds 247,342,127 (36GB) records in the File table, and 78,576,199 (10GB) records in the Filename table. Since 3.0.0, but even more since 3.0.1, I have a problem with queries being really slow. Basically when doing a full backup of a server (mailserver, LOTS of small files), I can have my MySQL hanging for up to 24+ hours on queries like this: INSERT INTO Filename( Name ) SELECT a.Name FROM ( SELECT DISTINCT Name FROM batch ) AS a WHERE NOT EXISTS ( SELECT Name FROM Filename AS f WHERE f.Name = a.Name ) with the status Sending data, and a lot of other similar queries in queue with the status Locked. One of these queries take approx 10.000 seconds to execute, but is just followed by another similar (identical) query with the same duration. This is a problem mostly because it prevents me doing restores while backups are Dir inserting attributes Obviously I would think it's a MySQL performance issue, but I was wondering if anything had been done to the queries? They seem to be a LOT slower and a LOT heavier. I've just put more RAM into the server, but it's done little to improve the duration of the queries. My server now has 4GB RAM (will update to 6GB) - but again it's a recent issue, because Bacula has been running perfectly for many months on 2GB RAM, until I updated to 3.0.1 I've done REPAIR TABLE and OPTIMIZE TABLE and seen no improvement. Finally if anyone have any specific ideas to improve performance on my huge SQL database, please share :) Sorry, I don't know how to optimize it. These queries are part of the new batch insert code. You can turn that off by passing --disable-batch-insert to configure when building Bacula. __Martin -- Crystal Reports - New Free Runtime and 30 Day Trial Check out the new simplified licensing option that enables unlimited royalty-free distribution of the report engine for externally facing server and web deployment. http://p.sf.net/sfu/businessobjects ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] MySQL errors
On Tue, 16 Jun 2009 14:29:25 +0100, John Kennedy said: We keep getting the following errors when running backups and any time we query: 16-Jun 13:56 ceres-dir JobId 8703: Fatal error: sql_get.c:359 sql_get.c:359 query SELECT VolumeName,MAX(VolIndex) FROM JobMedia,Media WHERE JobMedia.JobId=8703 AND JobMedia.MediaId=Media.MediaId GROUP BY VolumeName ORDER BY 2 ASC failed: Can't create/write to file '/tmp/#sql_555_0.MYI' (Errcode: 30) I can not find any problems with MySQL and /tmp is not full nor are there any permission issues. I am relatively new to Bacula and know a bit about MySQL (just short of Jr DBA level). Errcode 30 means Read-only file system so I think you might have problems with your /tmp directory. What are the permissions on it? Can you create files in /tmp from the shell when running as the same user as mysqld? Is it a separate partition and does mount show rw for it? Have you tried rebooting in single user mode and doing an fsck to check for errors? __Martin -- Crystal Reports - New Free Runtime and 30 Day Trial Check out the new simplified licensing option that enables unlimited royalty-free distribution of the report engine for externally facing server and web deployment. http://p.sf.net/sfu/businessobjects ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Performance with MySQL queries since 3.0.0 (Dir inserting attributes hang)
On Wed, 17 Jun 2009 13:48:58 +0200, Tom Sommer said: Martin Simmons wrote: On Tue, 16 Jun 2009 15:05:18 +0200, Tom Sommer said: Hi, I have a somewhat pressing problem with the performance of my Bacula installation. My MySQL database currently holds 247,342,127 (36GB) records in the File table, and 78,576,199 (10GB) records in the Filename table. Since 3.0.0, but even more since 3.0.1, I have a problem with queries being really slow. Basically when doing a full backup of a server (mailserver, LOTS of small files), I can have my MySQL hanging for up to 24+ hours on queries like this: INSERT INTO Filename( Name ) SELECT a.Name FROM ( SELECT DISTINCT Name FROM batch ) AS a WHERE NOT EXISTS ( SELECT Name FROM Filename AS f WHERE f.Name = a.Name ) with the status Sending data, and a lot of other similar queries in queue with the status Locked. One of these queries take approx 10.000 seconds to execute, but is just followed by another similar (identical) query with the same duration. This is a problem mostly because it prevents me doing restores while backups are Dir inserting attributes Obviously I would think it's a MySQL performance issue, but I was wondering if anything had been done to the queries? They seem to be a LOT slower and a LOT heavier. I've just put more RAM into the server, but it's done little to improve the duration of the queries. My server now has 4GB RAM (will update to 6GB) - but again it's a recent issue, because Bacula has been running perfectly for many months on 2GB RAM, until I updated to 3.0.1 I've done REPAIR TABLE and OPTIMIZE TABLE and seen no improvement. Finally if anyone have any specific ideas to improve performance on my huge SQL database, please share :) Sorry, I don't know how to optimize it. These queries are part of the new batch insert code. You can turn that off by passing --disable-batch-insert to configure when building Bacula. what does this mean in practice? 1 query per file? or? The non-batch approach makes several queries per file. My real problem is these huge queries lock the File table, so no other jobs can touch it for many hours. The batch queries update the Filename and Path tables atomically, so locking is expected. __Martin -- Crystal Reports - New Free Runtime and 30 Day Trial Check out the new simplified licensing option that enables unlimited royalty-free distribution of the report engine for externally facing server and web deployment. http://p.sf.net/sfu/businessobjects ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Performance with MySQL queries since 3.0.0 (Dir inserting attributes hang)
On Thu, 18 Jun 2009 17:11:04 +0200, Michel Meyers said: Martin Simmons wrote: On Wed, 17 Jun 2009 13:48:58 +0200, Tom Sommer said: Martin Simmons wrote: On Tue, 16 Jun 2009 15:05:18 +0200, Tom Sommer said: Hi, I have a somewhat pressing problem with the performance of my Bacula installation. My MySQL database currently holds 247,342,127 (36GB) records in the File table, and 78,576,199 (10GB) records in the Filename table. Since 3.0.0, but even more since 3.0.1, I have a problem with queries being really slow. Basically when doing a full backup of a server (mailserver, LOTS of small files), I can have my MySQL hanging for up to 24+ hours on queries like this: INSERT INTO Filename( Name ) SELECT a.Name FROM ( SELECT DISTINCT Name FROM batch ) AS a WHERE NOT EXISTS ( SELECT Name FROM Filename AS f WHERE f.Name = a.Name ) [...] Sorry, I don't know how to optimize it. These queries are part of the new batch insert code. You can turn that off by passing --disable-batch-insert to configure when building Bacula. what does this mean in practice? 1 query per file? or? The non-batch approach makes several queries per file. But it will do so as it backs up the files, as opposed to the batch approach where it collects all the queries and does them at the end of the job, right? Yes, but if you have spooling turned on, then it will spool the attributes too and insert them all at the end using several queries per file. __Martin -- Crystal Reports - New Free Runtime and 30 Day Trial Check out the new simplified licensing option that enables unlimited royalty-free distribution of the report engine for externally facing server and web deployment. http://p.sf.net/sfu/businessobjects ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Performance with MySQL queries since 3.0.0 (Dir inserting attributes hang)
On Fri, 19 Jun 2009 03:00:54 +0300, Jari Fredriksson said: INSERT INTO Filename( Name ) SELECT a.Name FROM ( SELECT DISTINCT Name FROM batch ) AS a WHERE NOT EXISTS ( SELECT Name FROM Filename AS f WHERE f.Name = a.Name ) That looks silly. I would write it shorter as INSERT INTO Filename(Name) SELECT DISTINCT Name FROM batch AS a WHERE NOT EXISTS ( SELECT Name FROM Filename AS f WHERE f.Name = a.Name ) Maybe, but perhaps it quicker to remove the duplicates from batch.name before comparing them to the rows in Filename.name? __Martin -- Crystal Reports - New Free Runtime and 30 Day Trial Check out the new simplified licensing option that enables unlimited royalty-free distribution of the report engine for externally facing server and web deployment. http://p.sf.net/sfu/businessobjects ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Performance with MySQL queries since 3.0.0 (Dir inserting attributes hang)
On Fri, 19 Jun 2009 09:51:20 +0200, Tom Sommer said: Martin Simmons wrote: On Thu, 18 Jun 2009 17:11:04 +0200, Michel Meyers said: Martin Simmons wrote: On Wed, 17 Jun 2009 13:48:58 +0200, Tom Sommer said: Martin Simmons wrote: On Tue, 16 Jun 2009 15:05:18 +0200, Tom Sommer said: Hi, I have a somewhat pressing problem with the performance of my Bacula installation. My MySQL database currently holds 247,342,127 (36GB) records in the File table, and 78,576,199 (10GB) records in the Filename table. Since 3.0.0, but even more since 3.0.1, I have a problem with queries being really slow. Basically when doing a full backup of a server (mailserver, LOTS of small files), I can have my MySQL hanging for up to 24+ hours on queries like this: INSERT INTO Filename( Name ) SELECT a.Name FROM ( SELECT DISTINCT Name FROM batch ) AS a WHERE NOT EXISTS ( SELECT Name FROM Filename AS f WHERE f.Name = a.Name ) [...] Sorry, I don't know how to optimize it. These queries are part of the new batch insert code. You can turn that off by passing --disable-batch-insert to configure when building Bacula. what does this mean in practice? 1 query per file? or? The non-batch approach makes several queries per file. But it will do so as it backs up the files, as opposed to the batch approach where it collects all the queries and does them at the end of the job, right? Yes, but if you have spooling turned on, then it will spool the attributes too and insert them all at the end using several queries per file. Well, the problem is really that first is spends time storing the files, and then it spends time saving information about the backup in the database. It would make more sense, for larger clients, to spend the idle database time (while taking the file backup) more wisely. I'm having a hard time judging if turning off batch-inserts is the right action, if I would gain anything from this. AFAIK, the non-batch approach in Bacula 3.x does the same queies as Bacula 2.x. Since you started the thread by saying that backups are slower in 3.x because of the queries, turning off batch-inserts might solve that. __Martin -- Crystal Reports - New Free Runtime and 30 Day Trial Check out the new simplified licensing option that enables unlimited royalty-free distribution of the report engine for externally facing server and web deployment. http://p.sf.net/sfu/businessobjects ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] [Bacula-devel] Heads up -- no Bacula encryption on FreeBSD Release 7
On Sun, 21 Jun 2009 11:55:43 -0400, Dan Langille said: Kern Sibbald wrote: Hello Dan, This is to warn you that Bacula will probably not be able to be compiled and run with encryption on Release 7 of FreeBSD. This is because the version of pthreads in that release has pthread_t defined as a structure, which is incompatible with OpenSSL. Initial conversations indicate that pthread_t is a point to a structure. From /usr/include/sys/_pthreadtypes.h typedef struct pthread *pthread_t; See http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/sys/_pthreadtypes.h Does this help? I concur. AFAIK, pthread_t is a pointer in all of the FreeBSD thread libraries so far. This will not work with the new FreeBSD implementation of pthreads. Perhaps the FreeBSD pthreads developers have a FreeBSD specific solution for this problem. If so, could you please let me know. If not, Bacula encryption will unfortunately not be available on the new version of FreeBSD. A FreeBSD project The FreeBSD port applies these thread related patches to the Bacula source: d...@polo:/usr/ports/sysutils/bacula-server/files] $ cat patch-src_lib_jcr.c --- src/lib/jcr.c.orig 2009-04-12 11:07:07.079940037 -0400 +++ src/lib/jcr.c 2009-04-12 11:07:20.116947254 -0400 @@ -1063,8 +1063,8 @@ fprintf(fp, \tuse_count=%i\n, jcr-use_count()); #else - fprintf(fp, \tuse_count=%i threadid=0x%x\n, - jcr-use_count(), (int)jcr-my_thread_id); + fprintf(fp, \tuse_count=%i threadid=0x%lx\n, + jcr-use_count(), (long)jcr-my_thread_id); #endif fprintf(fp, \tJobType=%c JobLevel=%c\n, jcr-get_JobType(), jcr-get_JobLevel()); [...@polo:/usr/ports/sysutils/bacula-server/files] $ --- src/cats/sql.c.orig 2009-04-12 11:07:46.218959381 -0400 +++ src/cats/sql.c 2009-04-12 11:08:15.509974021 -0400 @@ -759,7 +759,7 @@ if (mdb-lock.valid == RWLOCK_VALID) { fprintf(fp, \tRWLOCK=%p w_active=%i w_wait=%i\n, mdb-lock, mdb-lock.w_active, mdb-lock.w_wait); #ifndef HAVE_WIN32 - fprintf(fp, \t\tthreadid=0x%x mutex=%p\n, (int)mdb-lock.writer_id, mdb-lock.mutex); + fprintf(fp, \t\tthreadid=0x%lx mutex=%p\n, (long)mdb-lock.writer_id, mdb-lock.mutex); #endif } } These should be merged into Bacula, right? __Martin -- Are you an open source citizen? Join us for the Open Source Bridge conference! Portland, OR, June 17-19. Two days of sessions, one day of unconference: $250. Need another reason to go? 24-hour hacker lounge. Register today! http://ad.doubleclick.net/clk;215844324;13503038;v?http://opensourcebridge.org ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Restoring large directory does not work
On Tue, 23 Jun 2009 16:09:57 -0700, mehma sarja said: Did you wait till the cpu went back to low cpu usage? No, it stays high overnight and my patience runs out before cpu pegging does. I suggest attaching gdb to the bacula-dir process to see what it is doing, e.g. thread apply all bt Then detach gdb, let it run some more, and do the above again to see how it differs. You might need to build a debugging version of Bacula to get useful backtraces. __Martin -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Restoring large directory does not work
On Wed, 24 Jun 2009 10:41:03 -0700, mehma sarja said: (gdb) thread apply all bt Thread 4 (Thread 0x801902180 (LWP 100350)): #0 0x0008016f98cc in nanosleep () from /lib/libc.so.7 #1 0x0008009078c5 in nanosleep () from /lib/libthr.so.3 #2 0x0044e21e in bmicrosleep () #3 0x0042408d in wait_for_next_job () #4 0x00408a3c in main () Thread 3 (Thread 0x801902480 (LWP 100192)): #0 0x000801715afc in select () from /lib/libc.so.7 #1 0x0008009074d4 in select () from /lib/libthr.so.3 #2 0x0044f9b2 in bnet_thread_server () #3 0x00438ba8 in connect_thread () #4 0x000800908a27 in pthread_getprio () from /lib/libthr.so.3 #5 0x in ?? () Error accessing memory address 0x7fbff000: Bad address. #0 0x0040c043 in add_findex () The function add_findex is interesting, but I think like your bacula-dir was compiled without debugging info so it is difficult to see what is happening. Try the following gdb commands (I assume you are running 64-bit FreeBSD): break *add_findex commands printf arguments: %x %x %x\n, $rdi, $rsi, $rdx end continue This sets a breakpoint at the start of the function add_findex to print the arguments and starts Bacula running again. It should stop in gdb when it reaches the beginning of add_findex again (with a message like Breakpoint 1...in add_findex...). When it stops, enter the continue command again and time how long it takes before it stops again. Do this a few times and post the results (including the arguments: output). __Martin -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Restoring large directory does not work
On Wed, 24 Jun 2009 13:59:26 -0700, mehma sarja said: Thanks for all your help you guys. I am impressed with the level of expertise here! Error accessing memory address 0x7fbff000: Bad address. #0 0x0040c043 in add_findex () The function add_findex is interesting, but I think like your bacula-dir was Try the following gdb commands (I assume you are running 64-bit FreeBSD): break *add_findex commands printf arguments: %x %x %x\n, $rdi, $rsi, $rdx end continue When it stops, enter the continue command again and time how long it takes before it stops again. Do this a few times and post the results (including the arguments: output). Yes, it is FreeBSD 64 bit. The continue command comes right back with these arguments: Breakpoint 1, 0x0040bfc0 in add_findex () arguments: 1b17068 a0 5fe00b arguments: 1b17068 a0 5fe00b arguments: 1b17068 a0 5fe00b arguments: 1b17068 a0 5fe00b (gdb) continue Continuing. Breakpoint 1, 0x0040bfc0 in add_findex () arguments: 1b17068 a0 5fe039 arguments: 1b17068 a0 5fe039 arguments: 1b17068 a0 5fe039 arguments: 1b17068 a0 5fe039 (gdb) continue Continuing. Breakpoint 1, 0x0040bfc0 in add_findex () arguments: 1b17068 a0 5fe055 arguments: 1b17068 a0 5fe055 arguments: 1b17068 a0 5fe055 arguments: 1b17068 a0 5fe055 (gdb) continue Continuing. Breakpoint 1, 0x0040bfc0 in add_findex () arguments: 1b17068 a0 5fe060 arguments: 1b17068 a0 5fe060 arguments: 1b17068 a0 5fe060 arguments: 1b17068 a0 5fe060 (gdb) continue Continuing. Breakpoint 1, 0x0040bfc0 in add_findex () arguments: 1b17068 a0 5fe071 arguments: 1b17068 a0 5fe071 arguments: 1b17068 a0 5fe071 arguments: 1b17068 a0 5fe071 (gdb) continue Continuing. Breakpoint 1, 0x0040bfc0 in add_findex () arguments: 1b17068 a0 5fe079 arguments: 1b17068 a0 5fe079 arguments: 1b17068 a0 5fe079 arguments: 1b17068 a0 5fe079 (gdb) continue Continuing. Breakpoint 1, 0x0040bfc0 in add_findex () arguments: 1b17068 a0 5fe0ac arguments: 1b17068 a0 5fe0ac arguments: 1b17068 a0 5fe0ac arguments: 1b17068 a0 5fe0ac OK, this shows why it is slow. The algorithm in add_findex is only efficient when called with consecutive index values (the third number printed). The code for restore all in 2.4.4 doesn't do that, so it can take a very long time to complete. This was fixed in later version, so I think the best solution is to upgrade Bacula. __Martin -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Restoring large directory does not work
On Fri, 26 Jun 2009 11:53:26 -0700, mehma sarja said: Thanks Martin, You have put a good closure on the quest for knowledge. If I upgrade Bacula, will I have to upgrade the database? Meaning do I have to run those update table scripts. I am on postgresql version 8.29. Sorry, I don't know. Check the version table in the catalog. The latest Bacula uses version 11, so if your version table is the same then there should be no need to run update scripts. __Martin -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Tip: How to relabel an old tape without interrupting a running job that needs it
On Mon, 06 Jul 2009 16:37:00 -0400, Phil Stracchino said: Problem: You have a running job which turns out to need one volume more than you anticipated. You don't have another labelled tape ready. You DO have some available tapes waiting to be relabelled, but they're volumes with a different label scheme from an earlier installation. You can't use the label command because the tapes are already labelled. You can't use btape to label them because the sd has your tape drive locked, and you can't release the tape drive without stopping the sd, which will interrupt the job. Did you try the bconsole umount command? That releases the drive for me without stopping th sd. __Martin -- Enter the BlackBerry Developer Challenge This is your chance to win up to $100,000 in prizes! For a limited time, vendors submitting new applications to BlackBerry App World(TM) will have the opportunity to enter the BlackBerry Developer Challenge. See full prize details at: http://p.sf.net/sfu/blackberry ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Bacula 2.4.4 backup error
On Tue, 07 Jul 2009 08:13:43 +0200, Thomas Kempf said: hello, i've got some problems with bacula 2.4.4. on Debian Lenny and postgresql 8.3. Basically I'm facing the problem with postgresql discarding the complete batch insert after a job because there are some invalid utf8 byte-sequences in it. There were already some discussions on this subject with the proposal to dump the db and to reload it with encoding SQL_ASCII which is what i did. This did not solve my problem. The backup of some machines still produces this annoying error. Now i have two questions. 1. Why does bacula send me messages with the backup status of ok for those machines instead of telling me at least, that the batch copy failed. IMO there should be at least a warning that attribute despooling failed. If this happens with the latest version of Bacula, then please make a bug report. 2. How can i solve that problem. There is nothing in the bacula logs and the db-log just tells me that the batch insert failed at line xyz. But as the attribute spool-file is deleted after the job, i can't look which file produces the error. Is it something which is adressed in bacula v3. If so, i would take the pain and upgrade by self-compiling the new bacula instead of taking the deb-packages. What is the exact error you see in the db logs? What does psql -l print? __Martin -- Enter the BlackBerry Developer Challenge This is your chance to win up to $100,000 in prizes! For a limited time, vendors submitting new applications to BlackBerry App World(TM) will have the opportunity to enter the BlackBerry Developer Challenge. See full prize details at: http://p.sf.net/sfu/Challenge ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Speed writing to tape drive
On Mon, 6 Jul 2009 10:14:09 -0700, Hayden Katzenellenbogen said: Hello, Thanks to a patch published two weeks ago I am finally making headway into the wonderful world of Bacula. I have a single machine right now with about 1.2T of data I am backing up. When I run the btape fill test I get write speeds of around 70MB/s when I run a full backup from the local machine I get about 20MB/s is there anyway to figure out where the bottle neck is? Are you using software compression (gzip) in Bacula? What do top and iostat 10 show when running a backup? You could try using tar to check the speed of your filesystem, e.g. time tar cf /dev/null /some/very/large/directory __Martin -- Enter the BlackBerry Developer Challenge This is your chance to win up to $100,000 in prizes! For a limited time, vendors submitting new applications to BlackBerry App World(TM) will have the opportunity to enter the BlackBerry Developer Challenge. See full prize details at: http://p.sf.net/sfu/Challenge ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] backup size is bigger that sum of size of files
On Mon, 6 Jul 2009 09:42:37 -0400, Victor Hugo dos Santos said: hello, today I receive a strange report from bacula (about my two proxys servers), see: = yesterday JobId: 13223 Job:proxy2.2009-07-05_00.05.41 Backup Level: Full FileSet:Linux-Proxy 2008-02-29 00:05:14 Scheduled time: 05-Jul-2009 00:05:00 Start time: 05-Jul-2009 03:40:54 End time: 05-Jul-2009 03:57:52 Elapsed time: 16 mins 58 secs FD Files Written: 2,275 SD Files Written: 2,275 FD Bytes Written: 6,381,070,677 (6.381 GB) SD Bytes Written: 6,381,342,582 (6.381 GB) === today == JobId: 13246 Job:proxy2.2009-07-06_00.05.40 Backup Level: Incremental, since=2009-07-05 03:40:54 FileSet:Linux-Proxy 2008-02-29 00:05:14 Scheduled time: 06-Jul-2009 00:05:00 Start time: 06-Jul-2009 00:19:57 End time: 06-Jul-2009 00:35:26 Elapsed time: 15 mins 29 secs FD Files Written: 431 SD Files Written: 431 FD Bytes Written: 6,296,577,452 (6.296 GB) SD Bytes Written: 6,296,630,147 (6.296 GB) = now.. my FileSet: = FileSet { Name = Linux-Proxy Include { Options { signature = SHA1; compression=GZIP5 } File = /home File = /etc File = /root File = /var/log/squid3 File = /var/log/dansguardian File = /var/spool/cron } } = and finally the size of my folders. = $ sudo du --max-depth=0 -h /home/ /etc/ /root/ /var/log/squid3/ /var/log/dansguardian/ /var/spool/cron/ -c 67M/home/ 156M /etc/ 36K/root/ 2.5G /var/log/squid3/ 43M /var/log/dansguardian/ 20K /var/spool/cron/ 2.8G total = can you view the problem ??.. my file system only have 2.8GB, but my backup have a a size that 6 GB. Maybe you have some sparse files? You could check this by looking for large files using something like this: sudo find /home /etc /root /var/log/squid3 /var/log/dansguardian /var/spool/cron -xdev -size +200 -ls You could also do this in bconsole to see which files are in the Incremental backup and check their sizes: list files jobid=13246 I assume one of those files is sparse. Bacula has a fileset option to handle sparse files. __Martin -- Enter the BlackBerry Developer Challenge This is your chance to win up to $100,000 in prizes! For a limited time, vendors submitting new applications to BlackBerry App World(TM) will have the opportunity to enter the BlackBerry Developer Challenge. See full prize details at: http://p.sf.net/sfu/Challenge ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] backup size is bigger that sum of size of files
On Wed, 8 Jul 2009 12:53:43 -0400, Victor Hugo dos Santos said: On Wed, Jul 8, 2009 at 12:26 PM, Martin Simmonsmar...@lispworks.com wrote: You could also do this in bconsole to see which files are in the Incremental backup and check their sizes: list files jobid=13246 mmm.. before of send my mail. I ran this commando to check files included on backup, but not show problem or extra files for this job. +++-+--+---+--+---+---+ | jobid | name | starttime | type | level | jobfiles | jobbytes | jobstatus | +++-+--+---+--+---+---+ | 13,246 | proxy2 | 2009-07-06 00:19:57 | B| I | 431 | 6,296,577,452 | T | +++-+--+---+--+---+---+ That is strange -- the list files command should print all 431 of the file names. Has Bacula pruned the database already? You could try using the shell command bls -v to list the contents of the media. That should print the stat.st_size of each file. __Martin -- Enter the BlackBerry Developer Challenge This is your chance to win up to $100,000 in prizes! For a limited time, vendors submitting new applications to BlackBerry App World(TM) will have the opportunity to enter the BlackBerry Developer Challenge. See full prize details at: http://p.sf.net/sfu/Challenge ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Problem with Verify Jobs
It fails because you have two different job names for InitCatalog and Catalog. The Verify Job option isn't used for Catalog verifies. You need to run the Catalog verify using the same job as the InitCatalog but with the Level=Catalog option to the run command. __Martin On Thu, 09 Jul 2009 09:20:27 +0200, Christoph Kaminski said: Christoph Kaminski schrieb: no one a hint for me? have started the InitCatalog Job as first, then the Verify Job but I get ever these errors :/ Greetz Hi all! Im new on the list. I have a problem with verify jobs: lilith-fd JobId 95: Could not stat /etc/exports: ERR=No such file or directory 07-Jul 09:39 lilith-dir JobId 95: Bacula lilith-dir 2.4.4 (28Dec08): 07-Jul-2009 09:39:12 Build: x86_64-pc-linux-gnu debian lenny/sid JobId: 95 Job:lilith_InitCatalog.2009-07-07_09.36.20.39 FileSet:Verify Set Verify Level: InitCatalog Client: lilith-fd Verify JobId: 0 Verify Job: lilith_Verify Start time: 07-Jul-2009 09:36:22 End time: 07-Jul-2009 09:39:12 Files Examined: 37,401 Non-fatal FD errors:1 FD termination status: OK Termination:Verify OK 07-Jul 09:39 lilith-dir JobId 95: Begin pruning Jobs. 07-Jul 09:39 lilith-dir JobId 95: No Jobs found to prune. 07-Jul 09:39 lilith-dir JobId 95: Begin pruning Files. 07-Jul 09:39 lilith-dir JobId 95: No Files found to prune. 07-Jul 09:39 lilith-dir JobId 95: End auto prune. 07-Jul 12:06 lilith-dir JobId 96: Fatal error: Unable to find JobId of previous InitCatalog Job. Please run a Verify with Level=InitCatalog before running the current Job. 07-Jul 12:06 lilith-dir JobId 96: Error: Bacula lilith-dir 2.4.4 (28Dec08): 07-Jul-2009 12:06:41 Build: x86_64-pc-linux-gnu debian lenny/sid JobId: 96 Job:lilith_Verify.2009-07-07_12.06.39.40 FileSet:Verify Set Verify Level: Catalog Client: lilith-fd Verify JobId: 0 Verify Job: lilith_InitCatalog Start time: 07-Jul-2009 12:06:41 End time: 07-Jul-2009 12:06:41 Files Examined: 0 Non-fatal FD errors:0 FD termination status: Termination:*** Verify Error *** the config for this jobs: Job { Name = lilith_InitCatalog Client = lilith-fd JobDefs = DefaultInitCatalog FileSet = Verify Set } Job { Name = lilith_Verify Client = lilith-fd JobDefs = DefaultVerifyJob FileSet = Verify Set Verify Job = lilith_InitCatalog } What Im doing wrong? Greetz -- Enter the BlackBerry Developer Challenge This is your chance to win up to $100,000 in prizes! For a limited time, vendors submitting new applications to BlackBerry App World(TM) will have the opportunity to enter the BlackBerry Developer Challenge. See full prize details at: http://p.sf.net/sfu/blackberry ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Enter the BlackBerry Developer Challenge This is your chance to win up to $100,000 in prizes! For a limited time, vendors submitting new applications to BlackBerry App World(TM) will have the opportunity to enter the BlackBerry Developer Challenge. See full prize details at: http://p.sf.net/sfu/Challenge ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Enter the BlackBerry Developer Challenge This is your chance to win up to $100,000 in prizes! For a limited time, vendors submitting new applications to BlackBerry App World(TM) will have the opportunity to enter the BlackBerry Developer Challenge. See full prize details at: http://p.sf.net/sfu/Challenge ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Problem with Verify Jobs
On Thu, 09 Jul 2009 12:15:52 +0200, Christoph Kaminski said: Martin Simmons schrieb: It fails because you have two different job names for InitCatalog and Catalog. The Verify Job option isn't used for Catalog verifies. You need to run the Catalog verify using the same job as the InitCatalog but with the Level=Catalog option to the run command. __Martin YEP it works now... but I dont understand the function of this conf option then: Verify Job: lilith_Verify Have read the doc but still dont understand it... Can someone explain it with a example? The Verify Job option is only used for verifying backups. I don't know why Catalog verify ignores is, but it may be a bug. OTOH, you never need two jobs for InitCatalog/Catalog so it is simpler to have only one. __Martin -- Enter the BlackBerry Developer Challenge This is your chance to win up to $100,000 in prizes! For a limited time, vendors submitting new applications to BlackBerry App World(TM) will have the opportunity to enter the BlackBerry Developer Challenge. See full prize details at: http://p.sf.net/sfu/Challenge ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] What to do with Error Disk Volumes
On Wed, 15 Jul 2009 13:27:31 -0400, Mingus Dew said: All, I have a disk volume with a status of Error 149 | ShopCart-FS_Full-0149 | Error | 1 | 97,820,348,233 | 22 |1,814,400 | 1 |0 | 0 | ShopCart_File | 2009-07-02 02:59:40 Just wondering what is the proper method to clear or recover from this error? The best thing is to find out why it happened, before deciding how to correct it. Check the Bacula logs. __Martin -- Enter the BlackBerry Developer Challenge This is your chance to win up to $100,000 in prizes! For a limited time, vendors submitting new applications to BlackBerry App World(TM) will have the opportunity to enter the BlackBerry Developer Challenge. See full prize details at: http://p.sf.net/sfu/Challenge ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Backuping 2000+ host with bacula?
On Thu, 16 Jul 2009 12:09:32 +0200, Marcin Radczuk said: Hi, I'm trying use bacula more than a month. I have more than 2000 hosts to backup. We want to do everyday backup. One Full backup per month and Incrementals every other day. Now I have 10 bacula-sd and one bacula-dir and I was add only 300 hosts to schedule. I automated hosts add to bacula-dir, and I have one config file per host. Every host have his own Job, Pool and Fileset in his config. Problem is that bacula can do one backup per bacula-sd. I use Maximum Concurrent Jobs = 900 but I have one Pool per host and bacula can't run another job on bacula-sd if other job is running. Is that mean that I must do one pool per sd? The concurrency in the sd is limited by the number of Device entries, because each device can only load one volume at a time. Therefore you can use more than one pool concurrenly per sd if you define more than one Device in the sd config. Each device should have its own Media Type, which is used for the volumes in the pool. See http://www.bacula.org/en/rel-manual/Basic_Volume_Management.html#SECTION00265 __Martin -- Enter the BlackBerry Developer Challenge This is your chance to win up to $100,000 in prizes! For a limited time, vendors submitting new applications to BlackBerry App World(TM) will have the opportunity to enter the BlackBerry Developer Challenge. See full prize details at: http://p.sf.net/sfu/Challenge ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Spool block too big
On Thu, 16 Jul 2009 16:28:28 -0400, teksupptom said: Hello, We've been intermittently having an issue with backups failing due to the error Spool block too big. It's happened exactly 10 times since 4/27/09. It generally happens during large backups (900GB+). The most recent error happened after the data had been spooled, and was being written to tape. These usually occur overnight so I don't always get to see what's going on, but this one happened during my normal shift. Prior to it happening I had noticed that the SD status showed the correct data spool file size, but showed 0 bytes for the attribute spool size. I double checked the directory we use to store the attribute spool file (same directory where the mail files are kept, but different from where we spool the data), and there was a 6GB+ attribute spool file for the job. Not sure if this is what other people are seeing when this error occurs, but I'm hoping it can help in tracking down the source. We're running Bacula 2.4.4 using PostGreSQL 7.4.19. I think this can only be caused by a bug, so it is probably a good idead to upgrade to Bacula 3.0.1 to see if that fixes it. Are you using simultaneous jobs? What values does the error give for the sizes of the blocks? __Martin -- Enter the BlackBerry Developer Challenge This is your chance to win up to $100,000 in prizes! For a limited time, vendors submitting new applications to BlackBerry App World(TM) will have the opportunity to enter the BlackBerry Developer Challenge. See full prize details at: http://p.sf.net/sfu/Challenge ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] building bacula storage daemon only
On Mon, 20 Jul 2009 01:34:53 -0500, Bobby Castleberry said: Hello all, I am testing Bacula out and running into a bit of confusion. From the documentation it appears that I can have the storage daemon and director installed on separate systems. Yes. When searching for documentation for just building the storage daemon alone it seems like either it's not possible or nobody is really doing it. When I attempt a configure using: ./configure --sbindir=/opt/bacula/bin/ --sysconfdir=/opt/bacula/etc/ --mandir=/opt/bacula/man/ --disable-build-dird --enable-largefile --disable-nls It complains about not having an sql library defined. Does the storage daemon have a direct connection to the sql server without going through the director? Is there something obvious that I'm missing with configuration options? It isn't possible, because the storage daemon makefile also makes bscan, which does need the database. You'll need to install sqlite or the MySQL client package on the storage daemon build machine. __Martin -- Enter the BlackBerry Developer Challenge This is your chance to win up to $100,000 in prizes! For a limited time, vendors submitting new applications to BlackBerry App World(TM) will have the opportunity to enter the BlackBerry Developer Challenge. See full prize details at: http://p.sf.net/sfu/Challenge ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Bacula Compression
On Sun, 19 Jul 2009 15:54:35 +0100, Gavin McCullagh said: Hi, On Sun, 19 Jul 2009, Bruno Friedmann wrote: Another useful feature might be a wildcard excludefromcompression directive so you could say: This might avoid wasting time trying to compress the unccompressible while gaining the compression ratio on those files which gain from it? Wouah, that would be nice, is this already working with recent bacula version ? Not that I've heard of. I was just thinking out loud. You can do it already using multiple Options clauses: Options { # Files that we don't want to compress. signature = MD5 wild = *.gz wild = *.Z wild = *.mp3 } Options { # Default clause to compress all other files signature = MD5 compression = GZIP } Remember to put all other options (such as signature) into both clauses. __Martin -- Enter the BlackBerry Developer Challenge This is your chance to win up to $100,000 in prizes! For a limited time, vendors submitting new applications to BlackBerry App World(TM) will have the opportunity to enter the BlackBerry Developer Challenge. See full prize details at: http://p.sf.net/sfu/Challenge ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] incremental from different job
On Mon, 20 Jul 2009 11:57:51 +0200, C Keschnat said: is it somehow possible to use a Full backup from a different job for an incremental backup of another job? No. I have jobs for clients which run under a normal schedule (differential daily, full weekly). Some Server need to be backed up hourly. What I'd like to do is use the full or differential backup from the night before and from there start the incremental backups. The thing is that at night a script to dump all mysql-databases is called by bacula. I want mysqldump to run only at night, during the day I backup the binary logs. I have a solution for this, which is to simply run mysqldump via cronjob before bacula starts its work, but if this is solvable with bacula, I'd prefer that way You could add %l to the RunScript Command, which will pass the backup level to the script as $1. __Martin -- Enter the BlackBerry Developer Challenge This is your chance to win up to $100,000 in prizes! For a limited time, vendors submitting new applications to BlackBerry App World(TM) will have the opportunity to enter the BlackBerry Developer Challenge. See full prize details at: http://p.sf.net/sfu/Challenge ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Strange problem with authentication
On Mon, 27 Jul 2009 16:10:35 -0300, Marcelo Terres said: Nobody have any idea of what could cause this error ? The authentication between the sd and fd uses a session key generated by the director, so it is unclear how this can fail. Try running the 3 daemons at debug level 200 and see if that gives useful info. __Martin Regards, Marcelo H. Terres mhter...@gmail.com ICQ: 6649932 MSN: mhter...@hotmail.com Jabber: mhter...@jabber.org http://twitter.com/mhterres http://identi.ca/mhterres http://mhterres.jaiku.com/ http://mundoopensource.blogspot.com/ http://offtopicsandfun.blogspot.com/ http://www.propus.com.br On Fri, Jul 17, 2009 at 4:58 PM, Marcelo Terres mhter...@gmail.com wrote: Hi, I've got an odd problem. Let me explain. I had several clients and all of them was working perfectly until the end of last month (when I take some days off). I returned last wednesday, and 3 clients are reporting this problem: 17-Jul 16:38 backup-dir JobId 9073: Start Backup JobId 9073, Job=app253.2009-07-17_16.38.24 17-Jul 16:38 backup-dir JobId 9073: Using Device TapeDriveLTO3 17-Jul 16:38 backup-sd JobId 9073: Fatal error: Incorrect authorization key from File daemon at client rejected. Please see http://www.bacula.org/rel-manual/faq.html#AuthorizationErrors for help. 17-Jul 16:38 backup-sd JobId 9073: Fatal error: Incorrect authorization key from File daemon at client rejected. Please see http://www.bacula.org/rel-manual/faq.html#AuthorizationErrors for help. 17-Jul 16:38 backup-sd JobId 9073: Fatal error: Unable to authenticate File daemon 17-Jul 16:38 app253-fd JobId 9073: Fatal error: Authorization key rejected by Storage daemon. Please see http://www.bacula.org/rel-manual/faq.html#AuthorizationErrors for help. 17-Jul 16:38 app253-fd JobId 9073: Fatal error: Failed to authenticate Storage daemon. 17-Jul 16:38 backup-dir JobId 9073: Fatal error: Socket error on Storage command: ERR=No data available If I ask the client status in bconsole everything is fine: Connecting to Client app253 at 10.0.253.2:9102 app253-fd Version: 2.4.4 (28 December 2008) i486-pc-linux-gnu debian 4.0 Daemon started 17-Jul-09 16:37, 0 Jobs run since started. Heap: heap=138,492 smbytes=10,262 max_bytes=11,648 bufs=56 max_bufs=77 Sizeof: boffset_t=8 size_t=4 debug=0 trace=0 Running Jobs: Director connected at: 17-Jul-09 16:52 No Jobs running. So, I discard the possibility of password mismatch. This 3 servers are behind and router (a linux server), but they were behind this server for months without backup errors. Anybody has some tips to me ? The only thing I know is that the router had some changes, but nothing looks different, cause the other applications are working fine. I tried to turn off the firewall in the router but the problem persist, so I discard the firewall too. I'm using bacula 2.4.3 in the server and 2.4.4 in this client (the other 2 servers use 1.38 and I had a lot of 1.38 clients ini the network without problems) Regards, Marcelo H. Terres mhter...@gmail.com ICQ: 6649932 MSN: mhter...@hotmail.com Jabber: mhter...@jabber.org http://twitter.com/mhterres http://identi.ca/mhterres http://mhterres.jaiku.com/ http://mundoopensource.blogspot.com/ http://offtopicsandfun.blogspot.com/ http://www.propus.com.br Sent from Porto Alegre, RS, Brazil -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Diferential Backup
On Tue, 28 Jul 2009 15:31:14 -0300, Luiz Gustavo P Tonello said: Hello guys, I'm a newbie in Bacula, but i configure this in my office. So, i have one ask for us on the Diferential Backup: Bacula make one Full backup for week (Sunday), and one Diferential backup for day. But, i delete an archive on Friday. So, I don't know in which day this file was exclude. I will have to look in all jobs? or, as this archive existing in the last full job, i will find this in the last job? Sorry my bad english, ok? And I wait that they have understood. You can find the job using bconsole commands: Option 2 or 7 in the restore command. Option 1 or 2 in the query command. __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] rejected Hello command
On Tue, 30 Jun 2009 10:58:26 -0400, Shawn said: The problem is the following: 30-Jun 10:31 skydro-dir JobId 0: Fatal error: File daemon at 10.6.4.230:9102 rejected Hello command This happens on Microsoft Windows Vista x64, Windows 2008 Server x64, and Windows 2003 x64. Version I am using on the client ends are win64bacula-3.0.1.exe The file server itself is an Ubuntu 64-bit, jaunty, server edition, with latest bacula-director package, using mysql locally for a DB. The error sounds like a version mismatch to me. What is the exact version of your bacula-director package? I think it must be 3.x to work with a 3.x file daemon. If you are running a 3.x director, then try setting debug level 200 and see what output that generates. In particular, the filed debug message should show what the file daemon returns when it rejects the hello. __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] empty storage resource used for restore jobs
On Wed, 05 Aug 2009 12:08:55 -0500, Brian Clarkson said: I've almost completed my upgrades from 1.36.3 to 2.4.1 but I've encountered one last problem. I don't know the answer to your question. but 2.4.1 is an old release and is not even the latest in the 2.4.x branch. Can you try 3.0.2? __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Catalog too big / not pruning?
On Thu, 6 Aug 2009 05:59:24 -0700, Jeremy Koppel said: We're running Postgresql 8.0.8; we can't currently update this machine (we'll have to move Bacula to a newer box when we have one available). Ran that query, and the top 4 do have very large numbers: relname | reltuples | relpages -+-+-- file| 3.28168e+07 | 592614 OK, that is 147 bytes per row, which is about what you would expect. file_fp_idx | 3.28168e+07 | 378580 file_jobid_idx | 3.28168e+07 | 368832 file_pkey | 3.28168e+07 | 364870 And running vacuumdb with the --analyze flag says: INFO: file: found 0 removable, 32828342 nonremovable row versions in 592867 pages DETAIL: 0 dead row versions cannot be removed yet. Nonremovable row versions range from 113 to 154 bytes long. ... I ran the full vacuum after that, and now it is down to 5.9GB, so I guess all those records really weren't taking up much space. Also, the indexes actually got bigger: relname | reltuples | relpages -+-+-- file| 3.28283e+07 | 592684 file_fp_idx | 3.28283e+07 |90029 file_jobid_idx | 3.28283e+07 |71896 file_pkey | 3.28283e+07 |71895 I read up on it and saw that this was expected behavior, and that running a reindex on the table should fix it. So I ran REINDEX TABLE file;, but that didn't have any effect. I'll do some looking into that today. Look again at the sizes -- they actually got 5x smaller! Initially they were very bloated compared to the table size. Also, I found the output from dbcheck curious. Of all the orphaned records it found, the file records were an even number: 10,000,000. It sort of seems like maybe dbcheck can only clean 10,000,000 records at a time. : ) So, I have just now started running it again, and so far it has found 0 bad Path records, 0 bad Filename records, etc., all 0 this time, until it got to File records, where it says again: Found 1000 File records, Deleting 1000 orphaned File records. Yes, I think there is a limit on the number of file records it will delete each time. __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] number of files mismatch question
On Wed, 05 Aug 2009 12:05:32 -0400, uhog-v9e4 said: Hi all, I am a new bacula user running 3.02 on Ubuntu 8.04 LTS. I am backing up about 7 clients(mix of windows and linux) to a DDS3 tape autochanger. After testing and configuring for a few days I went live with this on Sunday with a full backup of all clients and doing incrementals Mon-Fri. So the problem... Yesterday I had a power outage. Bacula was idle, but there was a volume mounted in the drive. Today when the incrementals fired, I got an error: Maybe the power outage caused the drive to corrupt the tape? 05-Aug 10:25 pendual-dir JobId 39: Start Backup JobId 39, Job=mrc-vm1-backup.2009-08-05_10.25.39_04 05-Aug 10:25 pendual-dir JobId 39: Using Device Drive-1 05-Aug 10:25 mrc-vm1-fd JobId 39: DIR and FD clocks differ by 12 seconds, FD automatically compensating. 05-Aug 10:25 pendual-sd JobId 39: Volume 90m_1_1 previously written, moving to end of data. 05-Aug 10:26 pendual-sd JobId 39: Error: Bacula cannot write on tape Volume 90m_1_1 because: The number of files mismatch! Volume=1 Catalog=14 05-Aug 10:26 pendual-sd JobId 39: Marking Volume 90m_1_1 in Error in Catalog. Were there any syslog messages at this time? It correctly loaded the next volume in the autochanger and did the backups but I'm not sure why I am getting this error? A bls on the volume shows 14 files, but does have an interesting entry at the end. # bls -j -V 90m_1_1 -c /etc/opt/bacula/conf/bacula-sd.conf Drive-1|tee 90m_1_1.out # grep JobId= 90m_1_1.out |grep -v Volume Record|wc -l 28 90m_1_1.out excerpt ... Begin Job Session Record: File:blk=13:1 SessId=3 SessTime=1249401525 JobId=31 Job=mrc6320-backup.2009-08-04_17.00.00_06 Date=04-Aug-2009 17:19:06 Level=I Type=B End Job Session Record: File:blk=13:579 SessId=3 SessTime=1249401525 JobId=31 Date=04-Aug-2009 17:24:18 Level=I Type=B Files=417 Bytes=37,303,402 Errors=0 Status=T 05-Aug 10:16 bls JobId 0: Error: block.c:1010 Read error on fd=3 at file:blk 14:0 on device Drive-1 (/dev/nst0). ERR=Input/output error. What is the TWO EOFs setting for the device in Bacula's sd config? Does this match the setting in the drive? __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Catalog too big / not pruning?
On Fri, 7 Aug 2009 07:55:08 -0700, Jeremy Koppel said: I ended up running dbcheck 3 more times. The first time got another 10,000,000, the second another 8,000,000+, and the 3rd was trivial. Running it a fourth time came up all 0s. Running another full vacuum got the DB size down to 597MB, which sounds right. Nice space saving! In theory Bacula will remove stale file records, but there was a bug in some version that could leave them. So, there has been a job in the crontab that runs the standard vacuumdb (not full), but, it does this while Bacula is running. For the past few days, while running the full vacuum, I shut it off as a precaution. Perhaps I should modify the script to shut down Bacula during the standard vacuum? Is this needed? It is probably better to avoid vacuuming while Bacula is updating the database, but there is no need to shut it down. You could run the vacuum from a Bacula admin job rather than cron if you want to control it better. I run it after the backups, so pruning has had a chance to remove records. __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Migration job marking destination volume used!
On Thu, 6 Aug 2009 14:17:00 -0600, Robert LeBlanc said: I'm attempting to migrate some back-ups from our Data Domain box to tape and am migrating our first set of jobs. Once it finished migrating the first job it marked the LTO3 tape as used with only 50GB written. I thought the migration jobs would leave the volume in Append until it was full or met other criteria (volume use duration, etc) was satisfied. We are using 2.4.4. Any help on fixing this for the rest of the jobs? Look in the Bacula log to see if it says *why* it was marked as used at that point. Also check the syslog for messages about the tape drive device. __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] number of files mismatch question
On Fri, 7 Aug 2009 23:33:37 -0400, Mike Chisholm said: Were there any syslog messages at this time? Yes Aug 5 10:13:07 pendual kernel: [ 7737.941074] st0: Sense Key : Medium Error [current] Aug 5 10:13:07 pendual kernel: [ 7737.941150] Info fld=0xfc00 Aug 5 10:13:07 pendual kernel: [ 7737.941161] st0: Add. Sense: Sequential positioning error This repeated 3 more times at Aug 5 10:14:04 Aug 5 10:16:24 Aug 5 10:16:24 Looks very much like a hardware problem (tape corrupted). What is the TWO EOFs setting for the device in Bacula's sd config? Does this match the setting in the drive? There is no setting in the SD config so I think the default is TWO EOF=no. I'm not sure how to check this on the drive itself. I have tried various st, mtx, tapeinfo, and smartctl commands but none of them return something that is obvious to me. Sorry, I don't know on Linux (on FreeBSD mt has a geteotmodel command). It is probably set to one EOF by default like Bacula. FWIW, the last two days of backups to the next tape have performed flawlessly so perhaps this is a bad piece of media. It just seems pretty coincidental that I had this problem after a power failure. I'm working on getting this system hooked to a UPS as well. I'm not sure it's worth a lot of effort to try to figure this out, I mainly wanted to make sure I wasn't missing something obvious. I am willing to continue digging in it if anyone is really curious, otherwise I think I will just let thing continue and see how it goes. Assuming there was just some glitch caused by loss of power, I would think the next time bacula goes to write to this media it will read the header and begin writing new data with a new EOF after the first volume, then just appending additional volumes and writing new EOF's after each one, correct? So any count problem will just get overwritten. If it is actually bad media I will probably hit it again but then I'll know and can just toss the tape. Yes, a purged and recycled tape will be written from the start next time. __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Fwd: Migration job marking destination volume used!
On Fri, 7 Aug 2009 13:53:01 -0600, Robert LeBlanc said: It is really weird, I marked it as append and then when it filled up the second tape it was using with multiple migration jobs, it tried to use the first tape again since it was in the append state. It gave the message 07-Aug 09:31 babacula-dir JobId 40116: Max configured use duration exceeded. Marking Volume 43L3 as Used. The very odd thing is that it recycled this tape before the first migration job. There is a use duration of 20 days for tapes in this pool. I don't believe that 1.5 hours exceeds that. Again, the second tape that is used did not have this problem and it was recycled as well. I've marked the tape append again to see if it happens when the thrid tape is full. This is very confusing. It does sound confusing. Yoy could check that the tape itself has the correct use duration (the pool's use duration is only used to set the value for the tape when it is added to the catalog). The llist media pool= command will show this. __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] disabling 'accurate' for a restore
On Thu, 13 Aug 2009 10:05:14 +1000, James Harper said: I do a full backup once a week, and then incremental backups for the rest of the week. MSSQL is configured to back up transaction logs every day and keep the backup for 3 days (by which time it would have been backed up). It's Thursday today and I want to restore the most recent full backup (Saturday night) and all the subsequent transaction logs, but of course because accurate backup is on I'm going to have to do multiple restores at various points - restore as at Monday night to get Saturday + Sunday logfiles, then restore at Wednesday night to get Monday + Tuesday logfiles, etc. Restoring as at last nights backup will only give me the logfiles that existed at the time. So why are you using accurate backups for this case? __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Bacula backups and restores
On Thu, 13 Aug 2009 06:13:11 -0400, tqz said: --===6257829277651962677== Hello all Lets say in the event that disaster strikes and all data on my server is lost for whatever reason. I have encrypted backups of the data on the server stored on tapes (on and off site) and NAS. After reinstalling my database management system, and then firing up bacula using copies I have of the director, client and storage daemon files how can I then read the catalog from the encrypted tape. Is there any way I can test this without overwriting the bacula database. Also is there a way to read files (the director, storage and daemon) off the encrypted storage and then restoring them using bacula command line tools or non bacula command line toolsÂ….or am I not making any sense?!Â… The command line tool bextract can read tapes, but it can't read encrypted backups, so you would need to dump the catalog unencrypted. That may be OK, because all the information within it is stored on the tapes unencrypted anyway (Bacula only encrypts the contents of the files, not the filenames etc). Alternatively, you could use the Write Bootstrap option in the catalog backup job and keep a copy of this file for use with the restore command in bconsole. You need to keep a copy of each backup, since it records the information about the tape. Second thing is that I am noticing that the size of the backups on my different storage devices are different. So for example the first backup occurs on the tape and the size of the backup is 14.09gb, then on my second storage device it will be 14.12gb and my third one is 14.16gb. They are all full backups ??? Any idea why there may be different sizes Maybe encryption is non-deterministic? __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] level: Since?
The Since level is not implemented and AFAIK noone is working on it. Try using the bscan program to put the Full backup into the catalog. __Martin On Tue, 25 Aug 2009 13:47:23 +0200, Mapper ict department said: Does anyone have an answer on this yet? The reason i ask is because exactly at the end of the -one before last- full backup the server running the director and storage deamon crashed due to a kernel panic. At that moment the runbeforejob (dumping the bacula database / bacula catalog from mysql) f rom our job that writes the catalog to tape was busy. It might sound weird (or not) since there was only being read (i guess that's the only thing happening with a dump) from the database; tables were found corrupt. Now the full backup was level F(ull) and ended with status T(erminated normally). We were also able to cd through the directories and files, mark files and restore them. The next incremental job wanted to run a full because there was no proper full backup job found. Because of time constraints we were not able to run a full backup again. If the level: since is what i think it is we could have run manual incremental jobs and set the date so that files are backuped from that one day only, every night until the next full backup comes around again. Only there is not asked for a date and when the job runs with LEVEL SINCE it fails. Error mentions something like: this level is not supported. It would be nice though if it were possible. Does somebody has this working already? With kind regards, Tony Mapper ict department i...@mapperlithography.com 20-8-2009 19:12 Hello, Can anyone tell me what the Since level is for? I get there by: - typing run - choosing a job - typing mod - selecting 1: level With kind regards, Tony -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] How to verify an old volume
On Tue, 25 Aug 2009 10:03:43 -0600 (MDT), James L Rogers said: Hi, I am running Bacula Version: 3.0.2 (18 July 2009) x86_64-unknown-linux-gnu redhat on CentOS release 5.3. I want to verify a volume that was written several jobs ago. The volume was written to a USB drive and I am afraid that the file is corrupted. The volume is a differential backup that consists of 18 jobs (one job per file set) from 2 different clients. I have found several setups via google for the Verify job resource but nothing with my exact situation. Here is my current job resource which only tried to verify the last backup (incremental) of fileset0k and not the previous differential backup of fileset0k. Job { Name = VerifyUSB Type = Verify Level = VolumeToCatalog Messages = Standard Pool = differential # Client=server1.prinmath.com-fd Client = server2.prinmath.com-fd FileSet = fileset0k Verify Job = fileset0k } Here are my questions: 1) How do I specify the VerifyJob resource if I want an older job than the last backup of the specified fileset? Apparently a job number does work here. There is no way to do that with a VerifyJob at the moment. It always verifies the last matching jobid that ran. 2) Do I have to set up 18 Verify Jobs and run each one to verify the entire volume or can I address the volume in it's entirety (remembering that two clients wrote a total of 18 jobs to the volume). 3) Can anyone point me to a solution that is similar to my situation? You could check the low level consistency of the volume by using the bls command line tool, which should check at least that every block matches its checksum. __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] How do Jobs Queue?
On Mon, 31 Aug 2009 13:59:29 -0400, teksupptom said: Martin Simmons wrote: I think they are considered in the order that the Job definitions appear in the bacula-dir.conf. To guarantee the order, it is best to use different priorities. I use priorities 1, 2, 3 etc for the backups and 201, 202, 203 etc for the corresponding verifies. __Martin Hi Martin, I wish it were that simple, unfortunately the testing I've done shows there is more involved than just the order. My nightly Incrementals are set for the same time, and they get scheduled in the order they appear in the file. Then the different verify jobs are scheduled at the same time, but the verifies for one job type get scheduled, then the verifies for the other job type. Also, the first gets scheduled in one order (first client to last), while the other gets scheduled in opposite (last client to first). It's like Bacula scans the job list from top to bottom, then back up, and back down, etc, until all jobs are queued for that day. I like your suggestion about priorities, but the way I read the instructions was that jobs of a lower priority would wait until all jobs of a higher priority finish before starting. Since we have two tape drives, and use one of them for our large filesets, there would be times where the first tape drive would sit idle for a day or two while waiting for all the higher priority jobs to complete. Am I misunderstanding the way priority is implemented? I don't run concurrent jobs, so I'm not sure. If you set Allow Mixed Priority = yes in all jobs, then they may run concurrently at different priorities. __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Show disabled jobs
On Fri, 4 Sep 2009 20:12:08 -0300, Daniel Bareiro said: I've a productive server with Debian GNU/Linux Etch for Sparc which has installed Bacula 1.38.11. Is there in this version some way to obtain a listing of disabled jobs? The show jobs command prints it. __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] allways full even if incremental
On Tue, 08 Sep 2009 14:05:19 +0200, Gerald Leier said: hello, i noticed some strange behaviour when backing up one of our linux hosts with bacula. bacula allways backs up everything. if i run it 3 times in a row it allways produces the same amount of files (Incremental backup) at first i thought it may be because of some timestamps missing.. but the mount options are the same on hosts that dont do a allways full never incremental even if incremental backup. any suggestions how to debug that, or any ideas what can cause that behaviour? bacula version is 3.0.1 the filesystem taht makes trouble: / type ext3 (rw,acl,user_xattr) kernel: 2.6.16.60 Does Bacula upgrade all of the jobs to Full or is it running them as Incremental? The log output shows this. Are you using the same fileset for working and broken hosts? __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Building just SD from source
On Wed, 9 Sep 2009 11:32:45 +, Joseph L Casale said: Is this possible to do so that mysql binaries and libraries are not needed? Alas not, because building the SD also builds the tools like bscan, which do need a database. However, you can build Bacula with sqlite instead of mysql as long as you only need the bacula-sd program. Sqlite can be built from source in the depkgs download. __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Building just SD from source
On Wed, 9 Sep 2009 14:45:18 +, Joseph L Casale said: Accept-Language: en-US Content-Language: en-US Alas not, because building the SD also builds the tools like bscan, which do need a database. However, you can build Bacula with sqlite instead of mysql as long as you only need the bacula-sd program. Sqlite can be built from source in the depkgs download. Thanks for that info! Is it feasible for me to just jumpstart a vm quickly and compile it as per the manual, then copy the directory to the server in question and install just the SD, or am I overcomplicating? That should work, but I think you will need to Bacula's shared libraries as well as bacula-sd. Running ldd bacula-sd should show which ones (ignore anything in /usr/lib). __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Device FileStorage with MediaType File requested by DIR not found in SD Device resources.
On Mon, 14 Sep 2009 09:35:41 +0200, Roger Meier said: Hi all We have some problem with bacula. I don't know where the Problem is, because all of our 7 hosts have the same configuration (only different include/exclude file list and hostname) All works fine now for some weeks, but now since 3-4 days we have the following problem, on some of our system (not all, some works): 14-Sep 01:05 lin-backup1-sd JobId 990: Fatal error: Device FileStorage with MediaType File requested by DIR not found in SD Device resources. 14-Sep 01:05 lin-backup1-dir JobId 990: Fatal error: Storage daemon didn't accept Device FileStorage because: 3924 Device FileStorage not in SD Device resources. 14-Sep 01:05 lin-backup1-dir JobId 990: Fatal error: Max wait time exceeded. Job canceled. What does status storage=File print, in particular the Device status part? We have setup a cancel after 4 hours, but this shoul'd not happen, because the run of the backup is max. 1-2 hours. We have one Backup Host with one DIR and hone SD Daemon. See config snipped below: bacula-dir.conf: # Definition of file storage device Storage { Name = File Address = lin-backup1.intra SDPort = 9103 Password = pw4backup-sd Device = FileStorage Media Type = File Maximum Concurrent Jobs = 20 } bacula-sd.conf: Device { Name = FileStorage Media Type = File Archive Device = /data/backup LabelMedia = yes; Random Access = Yes; AutomaticMount = yes; RemovableMedia = no; AlwaysOpen = no; Autochanger = no } I don't know where the Problem is , because this config works some weeks but now this Errors :( Maybe the SD read a different config file? __Martin -- Come build with us! The BlackBerryreg; Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9#45;12, 2009. Register now#33; http://p.sf.net/sfu/devconf ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] regular expressions in FileSets
On Fri, 11 Sep 2009 09:07:57 +0200, Juergen Koch said: Hi, I want to split my vhost-Dir in 2 Jobs. I use this in the FileSet: File = /var/www/vhosts/[a-m]* and File = /var/www/vhosts/[n-z]* The Job starts, but I receive this message: Could not stat /var/www/vhosts/[n-z]*: ERR=File or Directory not found But, I think, I can use regular expessions in FileSets. Can anybody help? See the /home/a* example in FileSet Examples http://www.bacula.org/3.0.x-manuals/en/install/install/Configuring_Director.html#SECTION0068 __Martin -- Come build with us! The BlackBerryreg; Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9#45;12, 2009. Register now#33; http://p.sf.net/sfu/devconf ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Want to backup pst files in vista in addition to desktop and documents
On Mon, 21 Sep 2009 15:08:15 -0700, Syn, Joonho said: Here is my new FileSet for Vista clients. It backs up these locations: Documents Desktop Default pst location for outlook Mozilla Firefox profiles Internet Explorer Favorites I tried to get rid of as much cruft as possible and I only used exclude statements since I found using includes to not be workable in any manner that made sense. As such, if many other programs are installed besides office and firefox there will be more cruft to get rid of but it should be a trivial matter to add exclude lines for those as well. If you have further refinements please share, especially if they make this list shorter! I'll post another one for XP as soon as I get one up and running. You might get some ideas on how to do it from this example: http://article.gmane.org/gmane.comp.sysutils.backup.bacula.devel/12576 __Martin -- Come build with us! The BlackBerryreg; Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9#45;12, 2009. Register now#33; http://p.sf.net/sfu/devconf ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Building bacula-fd on OpenBSD 4.5
On Sun, 11 Oct 2009 20:43:22 -0300, Daniel Bareiro said: I'm trying to compile Bacula 3.0.2 with the source code from the official site. For it I've dowloaded the bacula-3.0.2.tar.gz file. ... Making libbaccfg.la ... /tmp/bacula-3.0.2/libtool --silent --tag=CXX --mode=link /usr/bin/g++-o libbaccfg.la parse_conf$(DEFAULT_OBJECT_TYPE) res$(DEFAULT_OBJECT_TYPE) -export-dynamic -rpath /usr/lib -version-info 1:0:0 /bin/sh: DEFAULT_OBJECT_TYPE: not found /bin/sh: DEFAULT_OBJECT_TYPE: not found It is the same with FreeBSD 4.9 -- some BSD make implementations can't handle Bacula's makefiles. Try using gmake (GNU make). __Martin -- Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now! http://p.sf.net/sfu/devconference ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Building bacula-fd on OpenBSD 4.5
On Mon, 12 Oct 2009 11:12:58 -0300, Daniel Bareiro said: Hi, Martin. On Monday, 12 October 2009 11:36:38 +0100, Martin Simmons wrote: On Sun, 11 Oct 2009 20:43:22 -0300, Daniel Bareiro said: I'm trying to compile Bacula 3.0.2 with the source code from the official site. For it I've dowloaded the bacula-3.0.2.tar.gz file. ... Making libbaccfg.la ... /tmp/bacula-3.0.2/libtool --silent --tag=CXX --mode=link /usr/bin/g++ -o libbaccfg.la parse_conf$(DEFAULT_OBJECT_TYPE) res$(DEFAULT_OBJECT_TYPE) -export-dynamic -rpath /usr/lib -version-info 1:0:0 /bin/sh: DEFAULT_OBJECT_TYPE: not found /bin/sh: DEFAULT_OBJECT_TYPE: not found It is the same with FreeBSD 4.9 -- some BSD make implementations can't handle Bacula's makefiles. Try using gmake (GNU make). With gmake it seems that no longer appear the errors that I've mentioned before: Great. ... The make of console draws attention to me since I am only installing bacula-fd. I think the client-only install includes the console because you sometimes want to run that on the client machine too. Then I get another problem with 'make install'. It will be that I need to install autoconf? fugu:/tmp/bacula-3.0.2# gmake install ./autoconf/mkinstalldirs /usr/sbin gmake: ./autoconf/mkinstalldirs: Command not found gmake: *** [installdirs] Error 127 I tried installing it, but I continue obtaining the same message of error. Does ./autoconf/mkinstalldirs exist in your build directory? It is part of the Bacula source distribution. __Martin -- Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now! http://p.sf.net/sfu/devconference ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Building bacula-fd on OpenBSD 4.5
On Thu, 15 Oct 2009 07:34:57 -0300, Daniel Bareiro said: On Wednesday, 14 October 2009 17:38:59 +0100, Martin Simmons wrote: In the FileSet resource for this client I'm doing backup of all the /etc, but there are some files and directories which aren't able to be copied. I consider that this is because the process (running with bacula: bacula) cannot access to those files and directories by the restrictive permissions that these have. Normally the bacula-fd should be run as root. Only the bacula-dir and possibly the bacula-sd shluld be run as the bacula user. Perfect. Thanks for the indication, Martin. Do you think that it is appropriate to starting bacula-fd with the script in /etc/bacula/scripts using the following lines in the /etc/rc.local file? # DGB - 20091012 # Starting bacula-fd on boot if [ -x /etc/bacula/scripts/bacula-ctl-fd ]; then /etc/bacula/scripts/bacula-ctl-fd start fi There is some way to normally stop daemon during shutdown/reboot of the operating system? If OpenBSD uses /usr/local/etc/rc.d like FreeBSD, then it would be better to add a script there. Otherwise, /etc/rc.local would be fine for startup. __Martin -- Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now! http://p.sf.net/sfu/devconference ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Tape Volumes
This probably won't fix the problem here, but I suggest changing /dev/st0 to /dev/nst0. Using /dev/st0 can cause data loss if the tape rewinds without Bacula knowing. __Martin On Thu, 15 Oct 2009 09:30:33 -0300, Pedro Bordin Hoffmann said: Thanks for all your help. I think Im understanding a little more of bacula :P im in my 3 tape, and is asking me this: Device Fita (/dev/st0) is not open. Device is BLOCKED waiting to create a volume for: Pool:FullBackupsTapePool Media type: Tape What is that? Thanks 2009/10/14 terryc ter...@woa.com.au Pedro Bordin Hoffmann wrote: Hello! Im doing some backup that required 2, 3 tapes.. so my question is.. The label of first tape is monday. and when I put a second tape.. should I name it like monday-2 ? or something else? Or Monday again? Yes, you can label the second tape monday-2 or similar When will I make a second time this backup, should I need to label the tape again? Or just put the tape when the other fills up? Once you label the tapes, they retain that label forever in baculas view. Bacula is different from other backup programs. It keeps track of the tapes and their contents (until record expiry dates are met) and runs things from the tape it decideds it wants (oldest expired). If you are accustomed to a GFS backup system, then you need to set up a seperate pool for each level. If you are one of those places that backups up stuff on monday to a specific set of tapes on monday, then you need a pool of tapes for such an activity.(you also need to carefuly set expiry dates so those tapes are available when next needed (say five days for a week or 12 days for a fortnight). If you want to post to the list an outline of your backup policy, then I'm sure people will give you some guidance/suggestions. -- Terry Collins {:-)} Bicycles, Appropriate Technology, Natural Environment, Welding -- Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now! http://p.sf.net/sfu/devconference ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now! http://p.sf.net/sfu/devconference ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] bextract -- restore single occurrence of file
On Sat, 17 Oct 2009 04:34:52 -0700 (PDT), cby said: Jesper Jesper Krogh wrote: I fired up bextract (standard bacula restore is not appropriate in this case) which restored the first occurrence of the file as expected. However, instead of halting after restoring the file, bextract continued to the end of the tape and found a second occurrence with the same filename which it duly restored replacing the originally restored file. I suggest you open a bug-report/feature request on this.. this seems like a really nasty situation to get into if stuff really burns. Jesper Interestingly, I tried Martin's suggestion of using a bootstrap file. I defined two parameters: Volume name = volname and Count = 1. My understanding was that once the number of restored files hit the Count value bextract would stop. This wasn't the case. That may be a bug. I also used an include-list file when running bextract so there may be a clash when using -b and -i command-line options. I wasn't able to provide all the parameters to use just a bootstrap file. I think you can get the job-related parameters by running bls. __Martin -- Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now! http://p.sf.net/sfu/devconference ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] version 3
On Thu, 22 Oct 2009 03:23:11 -0400, glynd said: When the backup runs I keep getting this error: 22-Oct 09:09 glyn-laptop-fd JobId 161: Could not stat /var/lib/bacula/bin/bacula.sql: ERR=The system cannot find the path specified. The file exists and is owned by bacula.root and chmod is 664 Why would this happen? This is the backup of the catalog, BTW Do have the correct client in the job definition? It is looking for the file on glyn-laptop-fd. __Martin -- Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now! http://p.sf.net/sfu/devconference ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] version 3
On Fri, 23 Oct 2009 09:56:49 -0400, glynd said: Sorry forget this # Backup the catalog database (after the nightly save) Job { Name = SugarCatalog JobDefs = SugarJob ... Job { Name = CumulusCatalog JobDefs = CumulusJob ... Job { Name = ColynCatalog JobDefs = ColynJob ... Job { Name = GlynCatalog JobDefs = GlynJob ... Job { Name = TyphoonCatalog JobDefs = TyphoonJob ... Wow, you have 5 different catalogs?! That's unusual. Are they running on 5 different machines (specified via the JobDefs), which are all different from your Bacula Director machine? If so, then you'll either need to use ClientRunBeforeJob and ClientRunAfterJob to dump the backups on the correct machine or have variants of the make_catalog_backup script to connect to the correct machine. Don't forget that the catalog backup job is just a normal job that uses a normal Bacula FD. You have to ensure that the sql file is present on that Bacula FD. __Martin -- Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now! http://p.sf.net/sfu/devconference ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] losing directory permissions and ownership
On Mon, 26 Oct 2009 10:52:47 +0200, Avi Rozen said: Hi, I'm at my wits end here. I'd appreciate any feedback on this. I have the following directory /home/avrozen/temp/bacula-test/owned-by-root/owned-by-avrozen/owned-by-bacula/bacula-test with some files in it. I've specifically chowned the directories and files to different users and groups. The relevant director configuration is at the bottom of this message. Note the single File directive pointing to the directory above. When I restore the files all parent directories except for bacula-test lose their permissions and owner/group (all are set to 0751/drwxr-x--x root:root). I'd like to cherry pick directories to backup and have them restored with correct ownership and permissions... What am I doing wrong? Is this the expected behavior? Yes, this is the expected behavior. During a backup, Bacula looks inside the directories you give in the 'File =' lines and doesn't save any information about the parents. It is difficult to get the cherry picking behaviour, but I think it is possible. You need to start higher up the tree and use two Options clauses to select the files to include and exclude. The first Options clause must match only what you want to back up (the parent directories themselves plus the cherry picked files/directories, but not other contents of the parent directories). The second Options clause should use wild = * and Exclude=yes to remove everything else. __Martin -- Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now! http://p.sf.net/sfu/devconference ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] version 3
On Mon, 26 Oct 2009 02:17:49 -0400, glynd said: Martin, Cedric, Thanks for the input. I think my understanding as to how to set up Bacula is a bit off! The configuration is this: mistral Ubuntu machine runs the Director, fd and sd with a USB hard drive connected to back up to. mistral hosts sugarcrm and the database is backed up. cumulus is a NAS without a client and an nfs mount point allows it to be backed up. typhoon Ubuntu machine runs a client and is backed up using the typhoon-fd Both glyn-laptop and colyn-acer-7 are windows laptops and are running clients. Both are backed up during the daytime schedule using the file daemons. From what you are saying, I don't need multiple catalogs, am I right? Correct. All of the jobs can share a single catalog. I would appreciate a critique of the director.conf which I have pasted below, so I can get Bacula running smoothly. I hope you don't mind pasting such a large lump... At a quick glance, it looks OK except for the catalog backups. You already have a single catalog, which is good. To back it up, you need only one Catalog backup job, specifying Client = mistral-fd. __Martin -- Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now! http://p.sf.net/sfu/devconference ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] losing directory permissions and ownership
On Tue, 27 Oct 2009 00:22:37 +0200, Avi Rozen said: Martin Simmons wrote: I'd like to cherry pick directories to backup and have them restored with correct ownership and permissions... What am I doing wrong? Is this the expected behavior? Yes, this is the expected behavior. During a backup, Bacula looks inside the directories you give in the 'File =' lines and doesn't save any information about the parents. Would be nice if it did... It is difficult to get the cherry picking behaviour, but I think it is possible. You need to start higher up the tree and use two Options clauses to select the files to include and exclude. The first Options clause must match only what you want to back up (the parent directories themselves plus the cherry picked files/directories, but not other contents of the parent directories). The second Options clause should use wild = * and Exclude=yes to remove everything else. Thanks for the explanation and tips. I tried it and a few other file set configurations, based on the examples in [1], specifically the following, which is supposed to include only home directories of users whose usernames start with either 'a' or 'b': FileSet { Name = Full Set Include { Options { wilddir = /home/a* wilddir = /home/b* } Options { RegexDir = .* exclude = yes } File = /home } } But I can't get bacula to do what I want. I tried the following file set to backup files in my test directory, but it excludes everything except for files in the root directory: FileSet { Name = test-fileset Include { Options { wilddir = /home/avrozen/temp/bacula-test/owned-by-root/owned-by-avrozen/owned-by-bacula } Options { regexdir = .* Exclude = yes } File = / } } Any ideas? What am I missing? You need to match the parent directories themselves and then the contents of owned-by-bacula: FileSet { Name = test-fileset Include { Options { # Match all directories leading up to the cherry picked directory regexdir = ^/home$ regexdir = ^/home/avrozen$ regexdir = ^/home/avrozen/temp$ regexdir = ^/home/avrozen/temp/bacula-test$ regexdir = ^/home/avrozen/temp/bacula-test/owned-by-root$ regexdir = ^/home/avrozen/temp/bacula-test/owned-by-root/owned-by-avrozen$ # Match the cherry picked directory itself regexdir = ^/home/avrozen/temp/bacula-test/owned-by-root/owned-by-avrozen/owned-by-bacula$ # Match the contents of the cherry picked directory wild = /home/avrozen/temp/bacula-test/owned-by-root/owned-by-avrozen/owned-by-bacula/* } Options { # Skip everything else wild = * Exclude = yes } File = / } } Note the $ at the end of each regexdir to make it match only the directory and nothing else with that prefix. __Martin -- Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now! http://p.sf.net/sfu/devconference ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] losing directory permissions and ownership
On Wed, 28 Oct 2009 06:25:54 +0200, Avi Rozen said: Martin Simmons wrote: You need to match the parent directories themselves and then the contents of owned-by-bacula: FileSet { Name = test-fileset Include { Options { # Match all directories leading up to the cherry picked directory regexdir = ^/home$ regexdir = ^/home/avrozen$ regexdir = ^/home/avrozen/temp$ regexdir = ^/home/avrozen/temp/bacula-test$ regexdir = ^/home/avrozen/temp/bacula-test/owned-by-root$ regexdir = ^/home/avrozen/temp/bacula-test/owned-by-root/owned-by-avrozen$ # Match the cherry picked directory itself regexdir = ^/home/avrozen/temp/bacula-test/owned-by-root/owned-by-avrozen/owned-by-bacula$ # Match the contents of the cherry picked directory wild = /home/avrozen/temp/bacula-test/owned-by-root/owned-by-avrozen/owned-by-bacula/* } Options { # Skip everything else wild = * Exclude = yes } File = / } } Note the $ at the end of each regexdir to make it match only the directory and nothing else with that prefix. __Martin Works! Thanks! It also works if I replace regexdir with wilddir (and remove the ^ and $). Ah, yes, wilddir will suffice here. I was just overly cautious about wilddir matching a prefix, but that only happens with a trailing *, for example merging wilddir = /a/b/c/owned-by-bacula wild = /a/b/c/owned-by-bacula/* into one line like this: wild = /a/b/c/owned-by-bacula* will spuriously match /a/b/c/owned-by-bacula-but-unwanted. __Martin -- Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now! http://p.sf.net/sfu/devconference ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Files/directories missing in backup
On Thu, 29 Oct 2009 11:07:30 +0100, Kjartan Kvamme said: I have a backup job set up to back up among other things several Mercurial repositories. Yesterday, I accidentally overwrote one of the hgrc files, so I wanted to restore it from a backup, but then I discovered that the backup apparently contained none of the hgrc files from any of the repositories, even though the other content (most of it at least) of the directory was in the backup. When I decided to look into it a bit more today, I found that the new hgrc file I had created as a replacement for the one I accidentally overwrote had now been included in the nightly backup, even though no copies of the previous versions had been backed up. Also, a new repository that was created yesterday had its hgrc included in the backup as well. Every other repository created before yesterday were missing their hgrc files. On further examination, I also discovered that some repository directories were missing entirely from the backup, and attempting to restore one of the repositories reveals that lots of files are missing from within it as well. The directory structure is like this: /home/source/hg/repos/ - Contains all repositories - some of these are included in the backup (except hgrc) and some were missing entirely from the backup Bacula is currently configured to back up /etc, /var/lib/mysql and /home/source. A full backup was run initially, and incremental backups run nightly. The operating system is Ubuntu 9.04. What could be causing this problem? The problem will occur if the files appear to be older than the date of the full backup. You can check for that by running /usr/bin/stat on one of the missing files. If that is the problem, then it could happened if you used mv to move an old tree of files into the /home/source/hg/repos/ directory from somewhere else. The solution is to either touch the files (I don't know if Mercurial will be confused by that) or use the Accurate option in Bacula 3.0 to ensure that it tracks these old new files. __Martin -- Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now! http://p.sf.net/sfu/devconference ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] [Bacula-devel] hang immediately after Start UA server
On Mon, 2 Nov 2009 18:57:09 -0800, Jo Rhett said: On Oct 22, 2009, at 11:10 AM, Martin Simmons wrote: I recommend getting rid of that MySQL client installation and using the precompiled package for the mysql-client port, which doesn't have this problem. Version 5.0 is on the media and other versions are on the FreeBSD ftp site. I downloaded mysql51-client from the freebsd site, and tried it out. Results? 1. Using it with the bacula compiled against the old revision produced an immediate error at the exact same point, saying it couldn't find the pthread function. So yeah, bacula had bound to a pthread call inside mysql and the downloaded client didn't have it. Great! So now I just recompile bacula against this mysql and it should fall through to the system pthread implementation, right? Nope. /d/usr/ports/sysutils/bacula-server/work/bacula-3.0.3/libtool --silent --tag=CXX --mode=link /usr/bin/c++ -L/usr/local/lib -L../lib -L../ cats -L../findlib -o bacula-dir dird.o admin.o authenticate.o autoprune.o backup.o bsr.o catreq.o dir_plugins.o dird_conf.o expand.o fd_cmds.o getmsg.o inc_conf.o job.o jobq.o migrate.o mountreq.o msgchan.o next_vol.o newvol.o pythondir.o recycle.o restore.o run_conf.o scheduler.o ua_acl.o ua_cmds.o ua_dotcmds.o ua_query.o ua_input.o ua_label.o ua_output.o ua_prune.o ua_purge.o ua_restore.o ua_run.o ua_select.o ua_server.o ua_status.o ua_tree.o ua_update.o vbackup.o verify.o -lbacfind -lbacsql -lbacpy -lbaccfg -lbac -lm -L/ usr/local/lib/mysql -lmysqlclient_r -lz -lpthread -lintl -lwrap - lssl -lcrypto backup.o(.text+0x1691): In function `send_accurate_current_files(JCR*)': : undefined reference to `db_accurate_get_jobids(JCR*, B_DB*, JOB_DBR*, db_list_ctx*)' backup.o(.text+0x16d6): In function `send_accurate_current_files(JCR*)': : undefined reference to `db_list_handler(void*, int, char**)' ua_restore.o(.text+0x2076): In function `user_select_jobids_or_files(UAContext*, RESTORE_CTX*)': : undefined reference to `db_accurate_get_jobids(JCR*, B_DB*, JOB_DBR*, db_list_ctx*)' vbackup.o(.text+0xe95): In function `do_vbackup(JCR*)': : undefined reference to `db_accurate_get_jobids(JCR*, B_DB*, JOB_DBR*, db_list_ctx*)' /usr/local/lib/libbacsql.so: undefined reference to `__er...@libmysqlclient_16' *** Error code 1 Do you have the broken bacula package installed in /usr/local still (or at least is /usr/local/lib/libbacsql.so still there)? This build is erroneously trying to link against it, which looks like a bug in the Makefile (putting -L/usr/local/lib at the beginning). That should be fixable by uninstalling the old bacula package first. Each subdirectory below this also fails out with the same problem. So I got the latest port from freebsd 5.1.40 and compiled that, and compiled 3.0.3 against that and it compiles fine back we're back to the same problem. Sorry, I can't parse your last sentence. Is this success or failure? __Martin -- Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now! http://p.sf.net/sfu/devconference ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Restore job don't know in which storage is the volume
On Fri, 30 Oct 2009 16:13:58 -0300, Victor Hugo dos Santos said: ... If I select yes... bacula show this message: == 30-Oct 15:13 bacula-sd JobId 16787: Please mount Volume oracle-completo-0901 for: Job: Restaurar.2009-10-30_15.13.09 Storage: FileStorage (/var/bacula) Pool: Default Media type: File == because the volumen isn't on storage0, but yes on storage1. well.. IMHO, the problem is in this line: -- Storage not found, using Storage Storage0 from MediaType File. Bootstrap records written to /var/lib/bacula/bacula-dir.restore.2.bsr -- why bacula, don't know in which local/storage it saved the volume and try to use the storage0 for default ?? Bacula uses the media type to choose the storage. You need to configure different media types for each storage and then update the existing volumes to match. __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Verify Job: Warning: The following files are in the Catalog but not on the Volume(s):
On Tue, 10 Nov 2009 10:54:26 +0100, Ralf Gross said: Martin Simmons schrieb: On Tue, 3 Nov 2009 09:51:17 +0100, Ralf Gross said: bacula 3.0.2, psql, debian etch Every now and then I receive error mails about missing files from verify jobs where I can't find the problem. Does it report all files as missing in that case or is it some subset? Not all. What does the end of the email look like, i.e. the job report from Build OS: to Termination:? This verify job complained only about missing files in /usr. Ah, this is a good clue. What is the fileset definition? Maybe it includes the /usr directory twice? __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] command line editing
On Wed, 11 Nov 2009 09:28:13 +0100, Christoph Litauer said: Dear bacula users, another question: I wonder if I can configure or compile bconsole in a way that I can use history and/or command line editing in a linux/mac terminal? Now, even typing a backspace confuses my terminal ... Try building with the --disable-conio argument to configure. Then Bacula will use readline if you have it installed (on Linux you'll need readline-devel as well). __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Verify Job: Warning: The following files are in the Catalog but not on the Volume(s):
On Wed, 11 Nov 2009 14:56:33 +0100, Ralf Gross said: Martin Simmons schrieb: On Tue, 10 Nov 2009 10:54:26 +0100, Ralf Gross said: Martin Simmons schrieb: On Tue, 3 Nov 2009 09:51:17 +0100, Ralf Gross said: bacula 3.0.2, psql, debian etch Every now and then I receive error mails about missing files from verify jobs where I can't find the problem. Does it report all files as missing in that case or is it some subset? Not all. What does the end of the email look like, i.e. the job report from Build OS: to Termination:? This verify job complained only about missing files in /usr. Ah, this is a good clue. What is the fileset definition? Maybe it includes the /usr directory twice? The fileset looks ok to me. FileSet { Name = Client Ignore FileSet Changes = yes Include { Options { aclsupport = yes signature = MD5 } File = / File = /var File = /boot } Exclude { File = /media/* File = /lost+found/* File = /mnt/* File = /dev/* File = /sys/* File = /proc/* File = /tmp/* File = /.journal File = /.fsck } } Yes, that looks OK, except for the use of * in the Exclude clause. The File = lines cannot contain wildcards. In your previous message, I see that Bacula reported: Files Expected: 50,755 Files Examined: 50,755 Does the backup contain exactly 50,755 files or has it miscounted them? Are you accidentally running two verify jobs at the same time? __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] after building bat, location of binary varies
On Sun, 15 Nov 2009 23:45:43 -0500, Dan Langille said: Folks, I am finding that the location of the executable binary varies from one system to another. I am trying to find out why. The answer will help to improve the build and install process. Sometimes the binary is at: src/qt-console/bat If not there, it is at: src/qt-console/.libs/bat Within a given system, the location is always consistent. It is one of the above. Why the location varies, I do not know. The .libs directory is the default location when building with libtool (for Bacula shared libraries). __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] after building bat, location of binary varies
On Mon, 16 Nov 2009 22:32:56 -0500, Dan Langille said: Martin Simmons wrote: On Sun, 15 Nov 2009 23:45:43 -0500, Dan Langille said: Folks, I am finding that the location of the executable binary varies from one system to another. I am trying to find out why. The answer will help to improve the build and install process. Sometimes the binary is at: src/qt-console/bat If not there, it is at: src/qt-console/.libs/bat Within a given system, the location is always consistent. It is one of the above. Why the location varies, I do not know. The .libs directory is the default location when building with libtool (for Bacula shared libraries). This is interesting. Please, can you elaborate? The libtool utility is a wrapper around compiling/linking/install to deal with portability for shared library naming. In the build tree, it puts all shared libraries and executables into subdirectories which are called .libs by default. It also creates a shell script for each executable, which sets LD_LIBRARY_PATH appropriately to make it work in the build tree. Note that this is only in the build tree. During make install, it installs the real libraries and binaries. My guess is that that some are not being linked with shared libraries for some reason, so Bacula is not using libtool and the real executable is built in src/qt-console/bat. It isn't clear to me why the location of the binaries matters, unless the Makefile is broken. __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Error with DR backups
On Mon, 16 Nov 2009 09:29:07 -0500, DAve said: 16-Nov 08:25 director-dir: Allied-ex3.2009-11-16_01.00.02 Warning: Error updating job record. sql_update.c:194 Update problem: affected_rows=0 16-Nov 08:25 director-dir: Allied-ex3.2009-11-16_01.00.02 Warning: Error getting job record for stats: sql_get.c:293 No Job found for JobId 20947 I am at a loss to understand why. The volumes can be pruned almost immediately as the backup is only for DR purposes and each volume will be recycled each night. The only problem I see is that the client is paying for 60GB and the backups have begun using more than that amount, so volumes are being reused within the current backup. That seems a very likely reason, especially if you have set Purge Oldest Volume = yes. When Bacula purges a volume, it removes whole jobs, not just the info for that volume. __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] after building bat, location of binary varies
On Tue, 17 Nov 2009 22:45:51 -0500, Dan Langille said: Martin Simmons wrote: On Mon, 16 Nov 2009 22:32:56 -0500, Dan Langille said: Martin Simmons wrote: On Sun, 15 Nov 2009 23:45:43 -0500, Dan Langille said: Folks, I am finding that the location of the executable binary varies from one system to another. I am trying to find out why. The answer will help to improve the build and install process. Sometimes the binary is at: src/qt-console/bat If not there, it is at: src/qt-console/.libs/bat Within a given system, the location is always consistent. It is one of the above. Why the location varies, I do not know. The .libs directory is the default location when building with libtool (for Bacula shared libraries). This is interesting. Please, can you elaborate? The libtool utility is a wrapper around compiling/linking/install to deal with portability for shared library naming. In the build tree, it puts all shared libraries and executables into subdirectories which are called .libs by default. It also creates a shell script for each executable, which sets LD_LIBRARY_PATH appropriately to make it work in the build tree. Note that this is only in the build tree. During make install, it installs the real libraries and binaries. My guess is that that some are not being linked with shared libraries for some reason, so Bacula is not using libtool and the real executable is built in src/qt-console/bat. It isn't clear to me why the location of the binaries matters, unless the Makefile is broken. It matters because building the FreeBSD port/packages needs to know where the binary is. Without knowing, you can't install it or build it into a package. OK, but I don't understand why. Normally, a port's Makefile (or bsd.port.mk) installs the software by running make install with the software's Makefile. The installed location should never have a .libs directory (it should only exist in the build tree). If the installed location has a .libs directory, then the software's Makefile is broken. It then builds the package from the installed files, so again there should be no .libs directory. __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] after building bat, location of binary varies
On Wed, 18 Nov 2009 20:12:43 -0500, Dan Langille said: Martin Simmons wrote: On Tue, 17 Nov 2009 22:45:51 -0500, Dan Langille said: Martin Simmons wrote: On Mon, 16 Nov 2009 22:32:56 -0500, Dan Langille said: Martin Simmons wrote: On Sun, 15 Nov 2009 23:45:43 -0500, Dan Langille said: Folks, I am finding that the location of the executable binary varies from one system to another. I am trying to find out why. The answer will help to improve the build and install process. Sometimes the binary is at: src/qt-console/bat If not there, it is at: src/qt-console/.libs/bat Within a given system, the location is always consistent. It is one of the above. Why the location varies, I do not know. The .libs directory is the default location when building with libtool (for Bacula shared libraries). This is interesting. Please, can you elaborate? The libtool utility is a wrapper around compiling/linking/install to deal with portability for shared library naming. In the build tree, it puts all shared libraries and executables into subdirectories which are called .libs by default. It also creates a shell script for each executable, which sets LD_LIBRARY_PATH appropriately to make it work in the build tree. Note that this is only in the build tree. During make install, it installs the real libraries and binaries. My guess is that that some are not being linked with shared libraries for some reason, so Bacula is not using libtool and the real executable is built in src/qt-console/bat. It isn't clear to me why the location of the binaries matters, unless the Makefile is broken. It matters because building the FreeBSD port/packages needs to know where the binary is. Without knowing, you can't install it or build it into a package. OK, but I don't understand why. Normally, a port's Makefile (or bsd.port.mk) installs the software by running make install with the software's Makefile. The installed location should never have a .libs directory (it should only exist in the build tree). If the installed location has a .libs directory, then the software's Makefile is broken. Yes, that is the normal situation. However, for sysutils/bacula-bat, the bat file is installed by the port's Makefile. I do not know why. That's how it's done. Perhaps the port's Makefile can be changed to use Bacula's make install like bacula-bat.spec does for the rpms? That must be better than adding more hacks... __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Help prunning oldest jobs
On Thu, 19 Nov 2009 17:05:19 -0500, ReynierPM said: Hi every: I need to free some space in the HDD where I make backups. So I want to delete some oldest volumes. I run this: *status Status available for: 1: Director 2: Storage 3: Client 4: All Select daemon type for status (1-4): 3 Automatically selected Client: FD.10.128.50.4 Connecting to Client FD.10.128.50.4 at 10.128.50.4:9102 salvasprod_dc_hosting1-fd Version: 2.4.4 (28 December 2008) i486-pc-linux-gnu debian 5.0 Daemon started 03-Nov-09 09:10, 0 Jobs run since started. Heap: heap=135,168 smbytes=9,502 max_bytes=9,624 bufs=52 max_bufs=53 Sizeof: boffset_t=8 size_t=4 debug=0 trace=0 Running Jobs: Director connected at: 19-Nov-09 17:03 No Jobs running. Terminated Jobs: JobId LevelFiles Bytes Status FinishedName == 1755 Full 26,54537.86 M OK 06-Sep-09 02:10 SP_DC_Hosting1-FD 1857 Incr 11,32037.60 M OK 16-Sep-09 02:05 SP_DC_Hosting1-FD 1908 Incr7812.055 M OK 17-Sep-09 02:05 SP_DC_Hosting1-FD 2217 Incr 8,32821.84 M OK 23-Sep-09 03:23 SP_DC_Hosting1-FD 2321 Incr3923.323 M OK 25-Sep-09 02:04 SP_DC_Hosting1-FD 2630 Incr9002.499 M OK 01-Oct-09 03:23 SP_DC_Hosting1-FD 2685 Incr3112.885 M OK 02-Oct-09 03:30 SP_DC_Hosting1-FD 3201 Incr70827.98 M OK 12-Oct-09 03:40 SP_DC_Hosting1-FD 3254 Incr 1,7935.485 M OK 13-Oct-09 03:49 SP_DC_Hosting1-FD 3307 Incr5872.094 M OK 14-Oct-09 03:45 SP_DC_Hosting1-FD Now how I prune some of this jobs? Wich is the right command? The Terminated Jobs listing above is not useful for this. Use the list jobs command to see what is still in the catalog and the prune or purge commands to remove jobs. After doing that, if you want to delete the volumes from disk, then use the delete command to remove the volume from the catalog before deleting it from disk (e.g. with rm). __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Help prunning oldest jobs
On Fri, 20 Nov 2009 09:27:13 -0500, ReynierPM said: Martin Simmons wrote: The Terminated Jobs listing above is not useful for this. Use the list jobs command to see what is still in the catalog and the prune or purge commands to remove jobs. After doing that, if you want to delete the volumes from disk, then use the delete command to remove the volume from the catalog before deleting it from disk (e.g. with rm). I'm a bit confuse at this point. Wich are the steps to follow? I mean for example: 1. List media 2. Prune (prune expired records from catalog) 2. Purge (purge records from catalog) 3. Delete Is that right or I miss something? Sorry, there is no simple sequence of steps, because it depends on which jobs you want to keep. If you already know which jobs and volumes you want to delete, then use the delete volume command, which deletes all jobs on a volume and removes the volume from the catalog. Then use rm to delete the file. In other cases, you can use the prune volume command and then use list volumes to find any Purged volumes, which can be deleted. __Martin -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] [Bacula-devel] Sun's incompetence
On Tue, 1 Dec 2009 17:32:45 +0100, Kern Sibbald said: On Tuesday 01 December 2009 17:15:40 Phil Stracchino wrote: Kern Sibbald wrote: On Tuesday 01 December 2009 16:31:45 Robert Hartzell wrote: Kern Sibbald wrote: Recently I tried loading OpenSolaris in a VM, and I discovered that they somehow have remained frozen in time in 1998 and had not significantly improved their user interface -- a big disappointment for me. I ended up with a default system that had a Java user interface and a /home/kern that was read-only (can you imagine that by default your home directory is read-only?). I finally gave up. What do you mean by Java user interface? Opensolaris is currently using Gnome Desktop version 2.28.0 That surprised me too. It was a full Java user interface, and all I did was about 6 months ago install an OpenSolaris. It is possible that they gave me the choice between two user interfaces and I made the bad choice. Java Desktop is what Sun is calling their Gnome-based desktop, afaik. They do offer two interface choices out of the box: the Gnome Java Desktop, or CDE. Much though I detest Gnome, it's the more usable of the two. Hmmm. That sounds like what I saw. I vaguely remember trying CDE, but it was too different, so I chose the Gnome desktop, which like you I don't particularly like, but I can work with it. However, Gnome running under or implemented in Java makes it *much* slower and nothing like what I would call a normal Gnome desktop running under X Window. Did you try switching the theme? The default Blueprint theme can be very slow. __Martin -- Join us December 9, 2009 for the Red Hat Virtual Experience, a free event focused on virtualization and cloud computing. Attend in-depth sessions from your desk. Your couch. Anywhere. http://p.sf.net/sfu/redhat-sfdev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] bacula restore - Volume data error
On Wed, 2 Dec 2009 13:18:00 +0100, Radim Roska said: Hello, I still have problem I mentioned in previous email. I've checked one random file that fails to be restored: db_mysql_dumps_1123-2009.tar.gz -rw-r--r-- 1 root root 252444672 2009-11-23 02:02 /tmp/bacula-restore/var/db_backups/db_mysql_dumps_1123-2009.tar.gz -rw-r--r-- 1 root root 281037550 2009-11-23 02:02 /var/db_backups/db_mysql_dumps_1123-2009.tar.gz and the file is obviously backed up not complete. So restoration process fails ( tar xfv on incomplete archive - tar: Unexpected EOF in archive tar: Error is not recoverable: exiting now ) 01-Dec 23:17 host-sd JobId 343: Ready to read from volume Volumes0097 on device FileStorage (/backups/). 01-Dec 23:17 host-sd JobId 343: Forward spacing Volume Volumes0097 to file:block 0:198. 01-Dec 23:25 host-sd JobId 343: Error: block.c:318 Volume data error at 8:3859489557! Block checksum mismatch in block=3D592437 len=3D64512: calc=3Dedab1d9b blk=3D820517c2 01-Dec 23:24 client-fd JobId 343: Error: attribs.c:421 File size of restore= d file /home/tmp/var/db_backups/db_mysql_dumps_1123-2009.tar.gz not correct. Original 281037550, restored 252444672. do you have idea what could be reason for this ? As I said hdd is fine. Was Bacula running with spooling enabled? If so, then the spool disk might have been faulty. Was Bacula running multiple concurrent jobs? There were bugs with that in some versions of Bacula. Have you tried something like smartctl (from smartmontools) to see if the drive has recorded any errors in the past? __Martin -- Join us December 9, 2009 for the Red Hat Virtual Experience, a free event focused on virtualization and cloud computing. Attend in-depth sessions from your desk. Your couch. Anywhere. http://p.sf.net/sfu/redhat-sfdev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Crashing storage director. Need help getting trace.
On Mon, 07 Dec 2009 14:30:41 +0800, Jim Barber said: Hi all. I have a problem where every weekend (or more frequently) my storage daemon crashes. The crash is random, but is happening either while running VirtualFull jobs or Copy jobs. So far it hasn't crashed during regular incremental backups. I am running version 3.0.3 of the Bacula software. First of all I tried adding a '-d 200' to the arguments that start bacula-sd. This produced a lot of messages, but nothing unusual that I can see prior to the crash. The last few lines in this log look like so: vc-sd: mac.c:241-468 before write JobId=468 FI=363302 SessId=1 Strm=MD5 len=16 vc-sd: mac.c:241-468 before write JobId=468 FI=363303 SessId=1 Strm=UATTR len=104 vc-sd: mac.c:241-468 before write JobId=468 FI=363304 SessId=1 Strm=UATTR len=122 vc-sd: mac.c:241-468 before write JobId=468 FI=363305 SessId=1 Strm=UATTR len=77 vc-sd: mac.c:241-468 before write JobId=468 FI=363305 SessId=1 Strm=DATA len=4496 vc-sd: mac.c:241-468 before write JobId=468 FI=363305 SessId=1 Strm=MD5 len=16 So next I have been trying to get the btraceback program running. I am using Debian packages (self built based on the 3.0.2 Debian sources). These run the storage daemon under the bacula:tape user:group. So I modified the btraceback program to use sudo to run gdb. I also configured sudo to allow the bacula user to do so without being prompted for a password. I then modified the Debian sources so that packages with debugging symbols are produced. If I become the bacula user and run a test like so: /usr/sbin/btraceback /usr/sbin/bacula-sd $PID Where: $PID = the process ID of the bacula-sd process, then I get an email showing debugging information. So as far as I can tell the btraceback program should be working. I had another crash of the storage daemon after making the changes and no email was sent. Nor was a bacula-sd.9103.traceback file produced. So I can't send any useful information to try and track down why the storage daemon is so unstable. It was also unstable when using the 3.0.2 Debian package as well so I don't think it is my rebuild that is causing the issue. Although I feel 3.0.3 is more stable than 3.0.2 was, I still can't get a complete weeks cycle working without a crash. The /etc/init.d/bacula-sd script defines the PATH to be, PATH=/sbin:/bin:/usr/sbin:/usr/bin So /usr/sbin is in the PATH and so I'd imagine the program should be able to find the traceback program. Any ideas how I can get some useful information from the crash? Try doing it interactively by attaching gdb to the bacula-sd process before it crashes (run gdb /path/to/bacula-sd and then use gdb's attach command). Then use the commands in btraceback.gdb when it crashes. __Martin -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Migration Job and Next Pool configuration
On Tue, 08 Dec 2009 11:00:52 +0100, Daniel Kamm said: Dear all, I am considering to use migration jobs to set up a D2D2T backup scenario. I am actually considering to use pools like that: Migration Job Backup Job -- WEEKLY1 Tape Pool / Backup data -- DEFAULT Disk Pool -- WEEKLY2 Tape Pool \ -- WEEKLY3 Tape Pool As far as I see, Next Pool configuration can only be done in the Pool resource and can only be done once. If I need n tape pools, I also need to configure n disk pools. I expected to be able to override NextPool directive at the Schedule (like you can override Pool directive there to select different pools for weekly backups). That would make it possible to have one disk pool and n tape pools for migration. Am I correct, that this is not possible? It isn't possible using the normal approach, but see http://thread.gmane.org/gmane.comp.sysutils.backup.bacula.devel/14084 __Martin -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Create new volume vs. recycle volume (with config)
On Thu, 10 Dec 2009 17:09:59 -0600, Brian Clarkson said: Is there a way to paginate the response from these list commands? I have 37k entries in this pool ... Not within Bacula, but you can use the @output command in bconsole: @output /tmp/out.txt list media pool=Default @output and then open /tmp/out.txt in a text editor. You can also run bconsole in a GNU Emacs shell (as long as bconsole is built without conio). __Martin On Thu, 2009-12-10 at 17:38 -0500, John Drescher wrote: On Thu, Dec 10, 2009 at 5:09 PM, Brian Clarkson bri...@lnstar.com wrote: On Thu, 2009-12-10 at 16:11 -0500, John Drescher wrote: On Thu, Dec 10, 2009 at 3:53 PM, Brian Clarkson bri...@lnstar.com wrote: On Thu, 2009-12-10 at 15:06 -0500, John Drescher wrote: Make the volume retention much longer. And then apply that to all of your volumes. Wouldn't the volume retention period need to be shorter? Retention period is how long to keep the volume before it can be reused. So if I don't want to reuse volumes I should set the retention period to zero? Definitely not. I know telling bacula to keep your backup data for 0 seconds is not what you want. Set this to the amount of time you want to keep your data after the backup is finished. Reusing the files (a volume is just a collection of 512m files per my configuration) is what I want to stop. I got you there. As you mentioned you need to turn off automatic volume recycling. Please post your configs. Have you had bacula reload the config file after you have changed it? I restarted the bacula director after making the changes. Subsequent backup jobs still re-used files. From bacula-dir.conf Pool { Name = Default Label Format = ${JobName}-${Level}-${NumVols} Pool Type = Backup Recycle = no AutoPrune = yes # Prune expired volumes Use Volume Once = yes Volume Retention = 25 days MaximumVolumeBytes = 512m Maximum Volume Jobs = 1 } Client { Name = client-fd Address = server.com FDPort = 9102 Catalog = MyCatalog Password = *** File Retention = 21 days Job Retention = 21 days AutoPrune = yes } Job { Name = Backup Server Type = Backup Client = server-fd FileSet = Full Set Schedule = WeeklyCycle Storage = Storage Messages = Standard Pool = Default Write Bootstrap = /usr/local/bacula/var/bacula/working/server.bsr } Storage { Name = Storage Address = storage.server.com SDPort = 9103 Password = *** Device = StorageServer Media Type = File } From bacula-sd.conf Device { Name = StorageServer Media Type = File Archive Device = /home/bacula-data/ LabelMedia = yes; # lets Bacula label unlabeled media Random Access = Yes; AutomaticMount = yes; # when device opened, read it RemovableMedia = no; AlwaysOpen = no; } I think I see why this is happening. Do the following list media pool=Default and see if Recycle=1 is set in any of the volumes. if so update and then select Pool from resource and follow the prompts. John -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] prune not working as expected
On Thu, 10 Dec 2009 13:11:38 -0800, Kevin Keane said: I have a couple older volumes that I expected to get pruned by now, but they don't. I am using bacula 3.0.2 on OpenSuSE. Here is one example: | 244 | randamai_20091101223426_Full.bacula| Archive | 1 | 93,005,712,230 | 21 |2,851,200 | 0 |0 | 0 | File | 2009-11-02 02:11:00 | When I used the prune volume command, I saw the following (all as expected): Enter *MediaId or Volume name: *244 The current Volume retention period is: 1 month 3 days Continue? (yes/mod/no): yes But this volume remains unpruned even though it is older than 1 month and 3 days. One thing that might possibly be related: these volumes were created with bacula 2.4.2. When I migrated to 3.0, I created a new database from scratch and used bscan to import the volumes. That was about a week ago. Bacula falsely assigned a retention time of 1 year; I corrected that with an update pool command. The problem is that the volume has status Archive, which was probably set by bscan. You can only prune volumes with status Full or Used. __Martin -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Continue to spool to disk when tape is full?
On Thu, 10 Dec 2009 15:51:13 -0500, Glen Barber said: Hello, I'm running Bacula version 2.4.3 on a machine with a single tape drive, and on occasion the tape will become full overnight. 'status dir' shows the following: server1.-MM-DD_HH.MM.SS is waiting on max Storage jobs Since the machine has plenty of space to spool the data, I'd like to continue to spool to disk when the tape is full so despooling to tape can take place when the tape is changed. I have the following Maximum Concurrent Jobs set: bacula-dir.conf: Maximum Concurrent Jobs = 1 bacula-sd.conf: Maximum Concurrent Jobs = 30 If I understand the documentation correctly, raising dir.conf Maximum Concurrent Jobs will interleave multiple backups to tape, which I want to avoid. Is there something obvious I've missed in the documentation, or is Bacula expected to stop spooling to disk when the tape is full? The latter -- it stops because it can't simultaneously read and write the spool file. __Martin -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] prune not working as expected
On Fri, 11 Dec 2009 13:30:52 -0800, Kevin Keane said: -Original Message- From: Martin Simmons [mailto:mar...@lispworks.com] Sent: Friday, December 11, 2009 12:44 PM To: bacula-users@lists.sourceforge.net Subject: Re: [Bacula-users] prune not working as expected On Thu, 10 Dec 2009 13:11:38 -0800, Kevin Keane said: I have a couple older volumes that I expected to get pruned by now, but they don't. I am using bacula 3.0.2 on OpenSuSE. Here is one example: | 244 | randamai_20091101223426_Full.bacula| Archive | 1 | 93,005,712,230 | 21 |2,851,200 | 0 |0 | 0 | File | 2009-11-02 02:11:00 | When I used the prune volume command, I saw the following (all as expected): Enter *MediaId or Volume name: *244 The current Volume retention period is: 1 month 3 days Continue? (yes/mod/no): yes But this volume remains unpruned even though it is older than 1 month and 3 days. One thing that might possibly be related: these volumes were created with bacula 2.4.2. When I migrated to 3.0, I created a new database from scratch and used bscan to import the volumes. That was about a week ago. Bacula falsely assigned a retention time of 1 year; I corrected that with an update pool command. The problem is that the volume has status Archive, which was probably set by bscan. You can only prune volumes with status Full or Used. Thanks! Yes, I discovered that right after I posted. But even after I updated the volume status to Used, prune still does not do anything. I also checked: there are no jobs associated with these volumes. Yes, prune should mark those volumes are purged if there are no jobs. It's a mystery. __Martin -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Crashing storage director. Need help getting trace.
On Mon, 14 Dec 2009 16:47:00 +0800, Jim Barber said: Jim Barber wrote: Thanks Martin. I've compiled and installed version 3.1.6 from a git pull I did on 10th Dec. I'm not sure if this new version will crash or not. But I've manually attached a gdb session to it just in case it does. Thanks. I'm not having much luck with this. When I attached to the process with gdb it seems to interfere with it. It's like to stops running. It no longer responds to status commands etc. I'm not familiar enough with gdb to resolve it. I tried the 'c'ontinue command just in case attaching stops the process. But it doesn't make any difference. Yes, you do need to use the continue command after attaching, but that should work. Possibly your version of gdb is broken, which might also explain the lack of email from btraceback. Did gdb print anything after you did that? It may be worth posting the whole gdb session. __Martin -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] prune not working as expected
On Tue, 15 Dec 2009 14:57:57 -0800, Kevin Keane said: Accept-Language: en-US Content-Language: en-US acceptlanguage: en-US -Original Message- I also checked: there are no jobs associated with these volumes. Yes, prune should mark those volumes are purged if there are no jobs. It's a mystery. Is there a way to turn on something like verbose output for pruning so I can find out exactly what it does to troubleshoot that? Running the Director at debug level 50 (or maybe 250) will print something about it, but I don't know if it will be enough. __Martin -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Block checksum mismatch error
On Fri, 18 Dec 2009 14:38:45 +0100, Oliver Lehmann said: Dan Langille writes: Have you considered that perhaps your backup is corrupted? That is, the USB2 HDD that you have backed up to contains a disk error? I would be very careful with that HDD, whatever you do, especially if that is your only backup of this data. I'm starting an offline SMART test now but I doubt that this will reveal anything. The disk has a UDMA_CRC_Error_Count of 1 but this is not that much. Right now I can't think that the disk has a problem. If this would be the case and I would access a file on a corrupt disk FreeBSD would have spammed my console already with read errors. Same goes for write. I'll let you know what the SMART analysis revealed. Is there a way to force bacula to re-read the backups after they where written or do I have to schedule continuous restores to make sure they stay readable? I would expect a verify job with level VolumeToCatalog to check the blocks. You need to run it immediately after the job, because it can only be used on the last run of each named job. __Martin -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Trouble on restore! Please help!
On Fri, 18 Dec 2009 18:01:57 +0100 (CET), Gabriele Bulfon said: Hello, I'm having trouble restoring some files from an old backup. The job is no more in my db, so I had to run a bscan to reconstruct it. I also tried with bextract, but because the backup is concurrent I got a wrong file. During bscan, here is a summary of the output: bscan: butil.c:282 Using device: QUANTUM for reading. 18-Dec 11:35 bscan: Invalid slot=0 defined in catalog for Volume GIOVEDI2 on QUANTUM (/dev/rmt/1cbn). Manual load my be required. 18-Dec 11:35 bscan: 3301 Issuing autochanger loaded? drive 0 command. 18-Dec 11:35 bscan: 3302 Autochanger loaded? drive 0, result is Slot 4. 18-Dec 11:35 bscan: Ready to read from volume GIOVEDI2 on device QUANTUM (/dev/rmt/1cbn). bscan: bscan.c:276 First Volume Size = 0nbscan: bscan.c:288 Using Database: bacula, User: bacula bscan: bscan.c:430 Pool record for Default found in DB. bscan: bscan.c:444 Pool type Backup is OK. bscan: bscan.c:454 Media record for GIOVEDI2 found in DB. bscan: bscan.c:469 VOL_LABEL: MediaType mismatch. DB=LIBRARY Vol=LTO bscan: bscan.c:976 Created Client record for Client: iserver-fd bscan: bscan.c:1059 Created new JobId=2344 record for original JobId=1936 18-Dec 11:35 bscan: End of file 1 on device QUANTUM (/dev/rmt/1cbn), Volume GIOVEDI2 bscan: bscan.c:681 32,768 file records. At file:blk=1:27,456 bytes=1,770,245,645 18-Dec 11:35 bscan: End of file 2 on device QUANTUM (/dev/rmt/1cbn), Volume GIOVEDI2 ... ... ... 18-Dec 13:42 bscan: End of file 223 on device QUANTUM (/dev/rmt/1cbn), Volume GIOVEDI2 18-Dec 13:42 bscan: End of Volume at file 223 on device QUANTUM (/dev/rmt/1cbn), Volume GIOVEDI2 bscan: bscan.c:318 == JobId=0 bscan: bscan.c:935 Updated Media record at end of Volume: GIOVEDI2 bscan: bscan.c:350 First Volume Size = 0n18-Dec 13:42 bscan: End of all volumes. bscan: bscan.c:935 Updated Media record at end of Volume: GIOVEDI2 bscan: bscan.c:630 End of all Volumes. VolFiles=223 VolBlocks=0 VolBytes=221,901,387,132 Records added or updated in the catalog: 1 Media 1 Pool 4 Job 1367418 File First, notice that bscan warned with a VOL_LABEL: MediaType mismatch. DB=LIBRARY Vol=LTO. I probably labeled these media when the MediaType was still LTO, then changed into LIBRARY (these are the names I use for single ltos and libraries). I hope this is not causing my problem. So, once finished, I ran bconsole, list jobs and found my job number. Checked that a list volumes was showing my current tapes to be in the slots correctly. Ran a restore via job number, marked 4 files in a directory, checked they were marked with lsmark and also with estimate, and it was correct. So I happily typed done to go on as usual...but I got this response: $ lsmark *simlifo *simlifo.14052009 *simlifo.311208 *simlifo.dopo6162 $ estimate 232387 total files; 4 marked to be restored; 417,792 bytes. $ done No files found to restore/migrate. No bootstrap file written. No files selected to be restored. What happened?! Anyone can help??? Which version are you running? Maybe it is fixed in 3.0.3? __Martin -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Trouble on restore! Please help!
I don't see any debug messages for that code, so I think the only way to debug it is using gdb on bacula-dir. The message No files found to restore/migrate. means that the function write_bsr has returned 0, which is unexpected. __Martin On Mon, 21 Dec 2009 09:26:27 +0100 (CET), Gabriele Bulfon said: Hi, it's 2.2.2, I can't upgrade yet. Is there any way I can check for errors? I can't find any inside the logs. Gabriele Bulfon - Sonicle S.r.l. Tel +39 028246016 Int. 30 - Fax +39 028243880 Via Felice Cavallotti 16 - 20089, Rozzano - Milano - ITALY http://www.sonicle.com -= Mail sent through WebTop2 =- -- Da: Martin Simmons A: bacula-users@lists.sourceforge.net Data: 18 dicembre 2009 22.56.30 CET Oggetto: Re: [Bacula-users] Trouble on restore! Please help! On Fri, 18 Dec 2009 18:01:57 +0100 (CET), Gabriele Bulfon said: Hello, I'm having trouble restoring some files from an old backup. The job is no more in my db, so I had to run a bscan to reconstruct it. I also tried with bextract, but because the backup is concurrent I got a wrong file. During bscan, here is a summary of the output: bscan: butil.c:282 Using device: QUANTUM for reading. 18-Dec 11:35 bscan: Invalid slot=0 defined in catalog for Volume GIOVEDI2 on QUANTUM (/dev/rmt/1cbn). Manual load my be required. 18-Dec 11:35 bscan: 3301 Issuing autochanger loaded? drive 0 command. 18-Dec 11:35 bscan: 3302 Autochanger loaded? drive 0, result is Slot 4. 18-Dec 11:35 bscan: Ready to read from volume GIOVEDI2 on device QUANTUM (/dev/rmt/1cbn). bscan: bscan.c:276 First Volume Size = 0nbscan: bscan.c:288 Using Database: bacula, User: bacula bscan: bscan.c:430 Pool record for Default found in DB. bscan: bscan.c:444 Pool type Backup is OK. bscan: bscan.c:454 Media record for GIOVEDI2 found in DB. bscan: bscan.c:469 VOL_LABEL: MediaType mismatch. DB=LIBRARY Vol=LTO bscan: bscan.c:976 Created Client record for Client: iserver-fd bscan: bscan.c:1059 Created new JobId=2344 record for original JobId=1936 18-Dec 11:35 bscan: End of file 1 on device QUANTUM (/dev/rmt/1cbn), Volume GIOVEDI2 bscan: bscan.c:681 32,768 file records. At file:blk=1:27,456 bytes=1,770,245,645 18-Dec 11:35 bscan: End of file 2 on device QUANTUM (/dev/rmt/1cbn), Volume GIOVEDI2 ... ... ... 18-Dec 13:42 bscan: End of file 223 on device QUANTUM (/dev/rmt/1cbn), Volume GIOVEDI2 18-Dec 13:42 bscan: End of Volume at file 223 on device QUANTUM (/dev/rmt/1cbn), Volume GIOVEDI2 bscan: bscan.c:318 == JobId=0 bscan: bscan.c:935 Updated Media record at end of Volume: GIOVEDI2 bscan: bscan.c:350 First Volume Size = 0n18-Dec 13:42 bscan: End of all volumes. bscan: bscan.c:935 Updated Media record at end of Volume: GIOVEDI2 bscan: bscan.c:630 End of all Volumes. VolFiles=223 VolBlocks=0 VolBytes=221,901,387,132 Records added or updated in the catalog: 1 Media 1 Pool 4 Job 1367418 File First, notice that bscan warned with a VOL_LABEL: MediaType mismatch. DB=LIBRARY Vol=LTO. I probably labeled these media when the MediaType was still LTO, then changed into LIBRARY (these are the names I use for single ltos and libraries). I hope this is not causing my problem. So, once finished, I ran bconsole, list jobs and found my job number. Checked that a list volumes was showing my current tapes to be in the slots correctly. Ran a restore via job number, marked 4 files in a directory, checked they were marked with lsmark and also with estimate, and it was correct. So I happily typed done to go on as usual...but I got this response: $ lsmark *simlifo *simlifo.14052009 *simlifo.311208 *simlifo.dopo6162 $ estimate 232387 total files; 4 marked to be restored; 417,792 bytes. $ done No files found to restore/migrate. No bootstrap file written. No files selected to be restored. What happened?! Anyone can help??? Which version are you running? Maybe it is fixed in 3.0.3? __Martin -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev
Re: [Bacula-users] Trouble on restore! Please help!
You can see the media type of the volumes with the list media command and change them using update media volume=... __Martin On Mon, 21 Dec 2009 12:42:01 +0100 (CET), Gabriele Bulfon said: hi, I tried to see how the bsr was written for a normal recent backup, and I find it like: Volume=SABATO MediaType=LIBRARY Device=QUANTUM Slot=6 VolSessionId=7 VolSessionTime=1261143945 VolFile=23 VolBlock=0-15498 FileIndex=3930 Count=1 I see MediaType is set to LIBRARY, as it is now. As you can see from my bscan, Bacula found a label with MediaType=LTO, because the tape was first labeled through a single lto device, that was originally named as MediaType=LTO, then the customer bought a Library device, and we changed configuration to MeditaType=LIBRARY, but we did not re-labeled tapes. Maybe bacula is causing an error because he cannot find the LTO MediaType? how can I dig into the DB for this LTO information? Maybe I can switch it to LIBRARY via sql? Thanks in advance. Gabriele. Gabriele Bulfon - Sonicle S.r.l. Tel +39 028246016 Int. 30 - Fax +39 028243880 Via Felice Cavallotti 16 - 20089, Rozzano - Milano - ITALY http://www.sonicle.com -- Da: Martin Simmons A: bacula-users@lists.sourceforge.net Data: 21 dicembre 2009 11.49.37 CET Oggetto: Re: [Bacula-users] Trouble on restore! Please help! I don't see any debug messages for that code, so I think the only way to debug it is using gdb on bacula-dir. The message No files found to restore/migrate. means that the function write_bsr has returned 0, which is unexpected. __Martin On Mon, 21 Dec 2009 09:26:27 +0100 (CET), Gabriele Bulfon said: Hi, it's 2.2.2, I can't upgrade yet. Is there any way I can check for errors? I can't find any inside the logs. Gabriele Bulfon - Sonicle S.r.l. Tel +39 028246016 Int. 30 - Fax +39 028243880 Via Felice Cavallotti 16 - 20089, Rozzano - Milano - ITALY http://www.sonicle.com -= Mail sent through WebTop2 =- -- Da: Martin Simmons A: bacula-users@lists.sourceforge.net Data: 18 dicembre 2009 22.56.30 CET Oggetto: Re: [Bacula-users] Trouble on restore! Please help! On Fri, 18 Dec 2009 18:01:57 +0100 (CET), Gabriele Bulfon said: Hello, I'm having trouble restoring some files from an old backup. The job is no more in my db, so I had to run a bscan to reconstruct it. I also tried with bextract, but because the backup is concurrent I got a wrong file. During bscan, here is a summary of the output: bscan: butil.c:282 Using device: QUANTUM for reading. 18-Dec 11:35 bscan: Invalid slot=0 defined in catalog for Volume GIOVEDI2 on QUANTUM (/dev/rmt/1cbn). Manual load my be required. 18-Dec 11:35 bscan: 3301 Issuing autochanger loaded? drive 0 command. 18-Dec 11:35 bscan: 3302 Autochanger loaded? drive 0, result is Slot 4. 18-Dec 11:35 bscan: Ready to read from volume GIOVEDI2 on device QUANTUM (/dev/rmt/1cbn). bscan: bscan.c:276 First Volume Size = 0nbscan: bscan.c:288 Using Database: bacula, User: bacula bscan: bscan.c:430 Pool record for Default found in DB. bscan: bscan.c:444 Pool type Backup is OK. bscan: bscan.c:454 Media record for GIOVEDI2 found in DB. bscan: bscan.c:469 VOL_LABEL: MediaType mismatch. DB=LIBRARY Vol=LTO bscan: bscan.c:976 Created Client record for Client: iserver-fd bscan: bscan.c:1059 Created new JobId=2344 record for original JobId=1936 18-Dec 11:35 bscan: End of file 1 on device QUANTUM (/dev/rmt/1cbn), Volume GIOVEDI2 bscan: bscan.c:681 32,768 file records. At file:blk=1:27,456 bytes=1,770,245,645 18-Dec 11:35 bscan: End of file 2 on device QUANTUM (/dev/rmt/1cbn), Volume GIOVEDI2 ... ... ... 18-Dec 13:42 bscan: End of file 223 on device QUANTUM (/dev/rmt/1cbn), Volume GIOVEDI2 18-Dec 13:42 bscan: End of Volume at file 223 on device QUANTUM (/dev/rmt/1cbn), Volume GIOVEDI2 bscan: bscan.c:318 == JobId=0 bscan: bscan.c:935 Updated Media record at end of Volume: GIOVEDI2 bscan: bscan.c:350 First Volume Size = 0n18-Dec 13:42 bscan: End of all volumes. bscan: bscan.c:935 Updated Media record at end of Volume: GIOVEDI2 bscan: bscan.c:630 End of all Volumes. VolFiles=223 VolBlocks=0 VolBytes=221,901,387,132 Records added or updated in the catalog: 1 Media 1 Pool 4 Job 1367418 File First, notice that bscan warned with a VOL_LABEL: MediaType mismatch. DB=LIBRARY Vol=LTO. I probably labeled these media when the MediaType was still LTO, then changed into LIBRARY (these are the names I use for single ltos and libraries). I hope this is not causing my problem. So, once finished, I ran bconsole, list jobs and found my job number. Checked that a list volumes was showing my current tapes to be in the slots correctly. Ran a restore via job number, marked 4 files in a directory, checked they were marked with lsmark and also with estimate
Re: [Bacula-users] Log partition weirdness
On Mon, 21 Dec 2009 16:27:49 -0700, Robert LeBlanc said: So, I've run into an interesting problem with 3.x. We have our servers configured for separate root, home and log partitions usually using LVM. I've seen this in the past, but now I'm looking to find an answer. We are running 3.0.2 on Debian Squeeze, when we include the /var/log/ partition, either by using oneFS on or oneFS yes and specifying it, it explodes the size of the backup. For instance, here is the layout of the partitions for one machine: FilesystemSize Used Avail Use% Mounted on /dev/mapper/lsgw-root 2.3G 842M 1.3G 39% / tmpfs 502M 0 502M 0% /lib/init/rw udev 10M 176K 9.9M 2% /dev tmpfs 502M 0 502M 0% /dev/shm /dev/sda1 228M 25M 192M 12% /boot /dev/mapper/lsgw-home 938M 18M 920M 2% /home /dev/mapper/lsgw-logs 3.6G 77M 3.6G 3% /var/log If we exclude the /var/log/ partition, estimate give us: 2000 OK estimate files=40539 bytes=711,837,359 When we include the /var/log/ partition using oneFS=no, estimate give us: 2000 OK estimate files=66741 bytes=233,544,784,662 When we include the /var/log/ partition using oneFS=yes, estimate gives us: 2000 OK estimate files=40620 bytes=232,347,795,189 df on the machine give us: du -sh / du: cannot access `/proc/4002/task/4002/fd/4': No such file or directory du: cannot access `/proc/4002/task/4002/fdinfo/4': No such file or directory du: cannot access `/proc/4002/fd/4': No such file or directory du: cannot access `/proc/4002/fdinfo/4': No such file or directory 800M/ As you can see a far cry from 232 GB. I've tried accurate backups without any change and the estimate from experience is what is actually backed up so it's not just a bad figure in the estimate command. Here is my fileset def: FileSet { Name = LinuxServer Include { Options { signature = MD5 onefs = yes exclude = yes wildfile = .reiserfs* wildfile = .journal wildfile = .autofsck wildfile = *~ } File = / File = /home File = /var/log } } For some reason /home is not affected by this, only /var/log. What can I do to try to track down what is going on, for now, I'll have to exclude /var/log, but I'd like to get to the bottom of this. Possibly you have sparse files in /var/log? Bacula can handle those better if you use the sparse=yes option in the FileSet (see the doc for more). __Martin -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Cannot restore VMmware/ZFS
On Tue, 29 Dec 2009 13:52:08 -0700, Paul Greidanus said: On 2009-12-28, at 6:56 PM, Marc Schiffbauer wrote: * Paul Greidanus schrieb am 28.12.09 um 23:44 Uhr: I'm trying to restore files I have backed up on the NFS server that I'm using to back VMware, but I'm getting similar errors to this every time I try to restore: 28-Dec 12:10 krikkit-dir JobId 1433: Start Restore Job Restore.2009-12-28_12.10.28_54 28-Dec 12:10 krikkit-dir JobId 1433: Using Device TL2000-1 28-Dec 12:10 krikkit-sd JobId 1433: 3307 Issuing autochanger unload slot 11, drive 0 command. 28-Dec 12:11 krikkit-sd JobId 1433: 3304 Issuing autochanger load slot 3, drive 0 command. 28-Dec 12:12 krikkit-sd JobId 1433: 3305 Autochanger load slot 3, drive 0, status is OK. 28-Dec 12:12 krikkit-sd JobId 1433: Ready to read from volume 09L4 on device TL2000-1 (/dev/rmt/0n). 28-Dec 12:12 krikkit-sd JobId 1433: Forward spacing Volume 09L4 to file:block 473:0. 28-Dec 12:15 krikkit-sd JobId 1433: Error: block.c:1010 Read error on fd=4 at file:blk 475:0 on device TL2000-1 (/dev/rmt/0n). ERR=I/O error. 28-Dec 12:15 krikkit-sd JobId 1433: End of Volume at file 475 on device TL2000-1 (/dev/rmt/0n), Volume 09L4 28-Dec 12:15 krikkit-sd JobId 1433: End of all volumes. 28-Dec 12:15 krikkit-fd JobId 1433: Error: attribs.c:423 File size of restored file /backupspool/rpool/vm2/.zfs/snapshot/backup/InvidiCA/InvidiCA-flat.vmdk not correct. Original 8589934592, restored 445841408. Files are backed up from a zfs snapshot which is created just before the backup starts. Every other file I am attempting to restore works just fine... Is anyone out there doing ZFS snapshots for VMware, or backing up NFS servers that have .vmdk files on it? No, but I could imagine that this might have something to do with some sparse-file setting. Have you checked how much space of your 8GB flat vmdk is aktually being used? Maybe this was 445841408 Bytes at backup time? Does the same happen if you do not use pre-allocated vmdk-disks? (Which is better anyway most of the times if you use NFS instead of vmfs) All I use is preallocated disks especially on NFS.. I don't think I can actually use sparse disks on NFS. As a test, I created a 100Gb file from /dev/zero, and tried backing that up and restoring it, and I get this: 29-Dec 13:45 krikkit-sd JobId 1446: Error: block.c:1010 Read error on fd=4 at file:blk 13:0 on device TL2000-1 (/dev/rmt/0n). ERR=I/O error. 29-Dec 13:45 krikkit-sd JobId 1446: End of Volume at file 13 on device TL2000-1 (/dev/rmt/0n), Volume 10L4 29-Dec 13:45 krikkit-sd JobId 1446: End of all volumes. 29-Dec 13:46 filer2-fd JobId 1446: Error: attribs.c:423 File size of restored file /scratch/rpool/vm2/.zfs/snapshot/backup/100GbTest not correct. Original 66365161472, restored 376340827. So, this tells me that whatever's going on, it's not Vmware that's causing me the troubles.. I'm wondering if I'm running into problems with ZFS snapshot backups, or just something with large files and Bacula? Looks like a tape drive or tape problem, since you get I/O error when reading. Did you check the syslog on the SD machine? Does bls give the same error for that volume? __Martin -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] No constraints in bacula database
On Wed, 06 Jan 2010 11:37:18 +0100, Marek Simon said: Hi, I do some experiments with bacula database and I found that there is no foreign key constraint between jobmedia and job, jobmedia and media, job and file and many similar logicaly connected tables. Why? Constraints prevent arising void references (File with empty job, File with empty filename, jobmedia record with empty job or media) I have bacula-dir version 2.4.4, postgres variant, and postgres 8.3, default instalation. Bacula inserts huge numbers of records into the File table during backups, so any constraints on that table could be very bad for performance. __Martin -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Is a Verify Job reliable on Windows ?
On Thu, 7 Jan 2010 21:29:55 +0100, Guillaume Laurès said: Hello fellow bacula users, I ran on a small issue today, using a Verify Job for the first time. I backed up 6000+ files, around 55 Gb (a Windows whole drive). Then I changed the letter of the drive, and restored all the files on a new partition under the original letter. Restore went okay, except for a minor a file number discrepancy (mark * trap ?). I need to wipe the original partition, and I don't remember if the restore job does a SHA1 check after restore against the Catalog, so I wanted to do one by myself before erasing the original data. Strangely enough running a Verify Job shows SHA1 differences for around 50 files :-( The oddity is that when I check-summed an original big zip file against the restored one with the file checksum utility from Microsoft, I found the very same SHA1 sig for both files. Note that this one is far from being equal to either one Bacula reports for this zip file... Some very small text files containing SHA1/MD5 cheksums where reported incorrect too, whereas they are indeed identical. Setup : - server is Ubuntu 8.04.3 LTS backup, with 3.0.3 installed from sources, virtualized in ESXi 3.5 (yes, I ran btape tests for several hours, ran fine albeit slower than drive official speed) - client is Windows 2003 Server 32bit, with 3.0.3a installed from the sourceforge binary - device is DLT-7000 with hardware compression turned off. Fileset used to backup : FileSet { Name = Donnees serveur Include { Options { Signature = SHA1 Compression = GZIP6 } File = E:/ } } The one used for the Verify Job (w/ Type = Disk to Catalog) just adds Verify = pinsug1 in the options under Compression. Any idea why Bacula thinks it's wrong whereas it's not ? I think the checksum includes everything returned by the BackupRead function on Windows, so possible something is slightly different in the meta data. __Martin -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] static bacula-fd for FreeBSD disaster recovery
On Sat, 9 Jan 2010 19:46:20 +0100, G VM said: Hi all, I am trying to build a static binary of bacula-fd. I found this thread: http://www.mail-archive.com/bacula-users@lists.sourceforge.net/msg33494.html http://sourceforge.net/mailarchive/message.php?msg_name=3838017D92B3C445BA200CE3B6A821A40A7482%40cetus.dawnsign.com I followed the instructions found there. I tried the way with putting the configure arguments in /etc/make.conf and with putting --enable-static-fd to the .configure line. Also with make static-bacula-fd I am using a fresh install of FreeBSD 7.2-RELEASE I see that configure is using this option. If I look in src/filed, I find a static-bacula-fd. However when I look at the linked libraries I see the same list as bacula-fd. Not 2 or 3 linked libraries. The version in ports is 3.x.x. Does anyone has the same experience? You need to configure with --disable-libtool to build static with 3.x.x. __Martin -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] bacula-dir error
On Tue, 12 Jan 2010 09:40:30 -0500, ReynierPM said: Hi every: Today I restart my Bacula Server for maintenance but I'm getting this error: 12-Jan 09:37 bacula-dir: ERROR TERMINATION at lex.c:270 Config token too long, file: /etc/bacula/bacula-dir.conf, line 23, begins at line 10 Why this? Mostly likely you've made a typo in the file. Look at line 10 for unbalanced quotes, since Bacula seems to think that is where a token starts. __Martin -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] problem restoring directory permissions
On Mon, 11 Jan 2010 11:06:04 -0800, Mark Nienberg said: Recently one of my users accidentally deleted some files and directories from a samba file server. I restored the missing stuff from a bacula backup, but the user could not access the restored files. Some troubleshooting revealed that the files were restored correctly, but the directories containing the files were restored with incorrect permissions. In order to understand what happened I did some simple testing on the bacula server itself. Here is the simple fileset for the server. FileSet { Name = khyber Files Include { Options { signature=MD5 compression=GZIP } File = /etc File = /home File = /root } # end Include } I created some subdirectories with various permissions inside the /root folder drwxrwxrwx 2 root root 4.1k Jan 8 17:17 testdir/ drwxr-xr-x 2 root root 4.1k Jan 8 17:12 testdir2/ drwxrws--- 2 root root 4.1k Jan 6 18:24 testdir3/ drwxr-xr-x 2 mark ftp 4.1k Jan 10 10:54 testdir4/ I added some test files inside the test folders. After a bacula backup was performed, I renamed the test directories by appending -save to them. Then I restored from the bacula backup and compared the restored directories to the originals. drwxrw-rw- 2 root root 4.1k Jan 11 10:31 testdir/ drwxrwxrwx 2 root root 4.1k Jan 8 17:17 testdir-save/ drwxr--r-- 2 root root 4.1k Jan 11 10:31 testdir2/ drwxr-xr-x 2 root root 4.1k Jan 8 17:12 testdir2-save/ drwxrw 2 root root 4.1k Jan 11 10:31 testdir3/ drwxrws--- 2 root root 4.1k Jan 6 18:24 testdir3-save/ drwxr--r-- 2 mark ftp 4.1k Jan 11 10:31 testdir4/ drwxr-xr-x 2 mark ftp 4.1k Jan 10 10:54 testdir4-save/ The problem seems to always be related to the group and others x permission. The files inside the test directories were all restored correctly. The root umask is 022, but that doesn't explain this. Note that bacula correctly restored the owner and group so it is obviously saving the information about the directories. Can anyone shed any light on what may be wrong with my configuration? Or can you confirm that the same happens on your system? Did you mark the directories to be restored or only the files? I think the problem could occur if you only restored the files. What was the Replace setting for the restore? At some point, restoring with Replace not set to always would cause problems with the directory permissions, but that might have been fixed. __Martin -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] static bacula-fd for FreeBSD disaster recovery
On Wed, 13 Jan 2010 15:01:28 +0100, G VM said: I have put this in /etc/make.conf:. if $(.CURDIR:M*/sysutils/bacula-client) || \ $(.CURDIR:M*/sysutils/bacula-client) CONFIGURE_ARGS+=--disable-libtool .endif When doing: make config I selected SFDAEMON. After compilation I got a static-bacula-fd of about 2MB. when I do: ldd static-bacula-fd ldd: static-bacula-fd: not a dynamic ELF executable I expected to see a few libraries. Are there absolutely no libraries needed now? Or is something wrong here? That is the definition of static -- libc etc are built into it, so there are no libraries needed. It is like /sbin/init and the files in /rescue. This command should show that it is statically linked: file static-bacula-fd __Martin -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] bacula-fd on windows: ... is a different filesystem. Will not descend from....
On Wed, 13 Jan 2010 17:13:57 +0100, Carlo Filippetto said: The same is with Linux, If I remember well.. somewhere I read something like ... it skips those files that are not 'valid' or file that the system recreates at startup You can try to set onefs=no I think that is a bad idea, because it will generate two separate copies of the files and restore will not work correctly. __Martin -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Bnet receive challenge response error ? Authentication errors...
On Wed, 13 Jan 2010 17:16:21 +0100, Simon Spring said: Hi all I do have a quite misterious and headacking problem with one of my fresh hosts, which i'd like to integrate with bacula. Whatever i try or do, i cannot connect to the bacula-fd on my client, i allways get the message below: * 13-Jan 16:57 godzilla-dir JobId 0: Fatal error: Unable to authenticate with File daemon at theclient.dom.local:9102. Possible causes: Passwords or names not the same or Maximum Concurrent Jobs exceeded on the FD or FD networking messed up (restart daemon). Please see http://www.bacula.org/rel-manual/ faq.html#AuthorizationErrors for help. ** What i am able: telnet to the host (by ip and fqdn) on the respective port 9102 without problems. I ensured at least 5 times (and let it countercheck by collegue), that passwords do match on both hosts/sides (yes, they do). There is no firewall active on both sides, hosts.allow allows anything, host and ip resolve (ping etc.) vicecersa correctly. Here is what i get from debug on the client: $ bacula-fd -c /etc/bacula/bacula-fd.conf -v -s -d350 -f bacula-fd: filed_conf.c:438-0 Inserting director res: godzilla-mon gnadmin-fd: message.c:259-0 Copy message resource 809be78 to 8099a70 gnadmin-fd: jcr.c:135-0 read_last_jobs seek to 188 gnadmin-fd: jcr.c:142-0 Read num_items=0 gnadmin-fd: pythonlib.c:113-0 No script dir. prog=FDStartUp gnadmin-fd: filed.c:225-0 filed: listening on port 9102 gnadmin-fd: bnet_server.c:96-0 Addresses host[ipv4:192.168.99.24:9102] gnadmin-fd: bnet.c:667-0 who=client host=192.168.99.10 port=36387 gnadmin-fd: jcr.c:617-0 OnEntry JobStatus=gnadmin-fd: jcr.c:637-0 OnExit JobStatus=C set=C gnadmin-fd: find.c:81-0 init_find_files ff=809c488 gnadmin-fd: job.c:233-0 dird: Hello Director godzilla-dir calling gnadmin-fd: job.c:249-0 Executing Hello command. gnadmin-fd: job.c:359-0 Calling Authenticate gnadmin-fd: cram-md5.c:73-0 send: auth cram-md5 696202879.1263398...@gnadmin-fd ssl=0 gnadmin-fd: cram-md5.c:89-0 Bnet receive challenge response error. This line is the problem -- there was an error receiving the reply for some reason. Unfortunately, Bacula doesn't print the error message. You could try attaching strace to the bacula-fd while connecting, to see if there is a errno reported in one of the socket syscalls. gnadmin-fd: authenticate.c:131-0 cram_auth failed for 192.168.3.10 gnadmin-fd: job.c:252-0 Quit command loop. Canceled=0 gnadmin-fd: runscript.c:103-0 runscript: running all RUNSCRIPT object (ClientAfterJob) JobStatus=C gnadmin-fd: pythonlib.c:237-0 No startup module. gnadmin-fd: job.c:343-0 Calling term_find_files gnadmin-fd: job.c:346-0 Done with term_find_files gnadmin-fd: mem_pool.c:377-0 garbage collect memory pool gnadmin-fd: job.c:348-0 Done with free_jcr 192.168.99.10 is bacula server deamon 192.168.99.24 is the client What is 192.168.3.10 (from the authenticate.c:131-0 line above)? Maybe you have routing problems? __Martin -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] SQL query
On Wed, 13 Jan 2010 14:18:18 +0100, David Dumortier said: Importance: Normal Hi all, I try to query MySQL to obtain all the date of backup for a file. I used this but have only the last backup : :List all backup for a File with date : *Enter Client : *Enter Filename (no path) : SELECT Client.Name as Client, Path.Path,Filename.Name, Job.JobId as JobId, StartTime FROM Client,Job,File,Filename,Path WHERE Client.ClientId=Job.ClientId AND Job.JobId=File.JobId AND Path.PathId=File.PathId AND Filename.FilenameId=File.FilenameId AND Client.Name='%1' AND Filename.Name='%2' ORDER BY Path.Path; That should work. Maybe your catalog has been pruned, so the file records are no longer there? __Martin -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Question about restores and multiple pools
On Tue, 12 Jan 2010 16:38:08 -0600, Jon Schewe said: I have standard backups (incremental, differential, full) going to tape using a pool named Default. I have weekly full backups going to an external drive using a pool named Offsite. Now when I do a restore I would like bacula to only consider the backups in the Default pool and to ignore those in the Offsite pool. Is this possible? No. Related question: When bacula is considering which is the most recent full backup to compute differences for a differential backup will it consider all job definitions and thus consider my Offsite pool? Or will it just consider just the currently running Job definition and thus use only the Default pool (tapes)? It will consider the most recent occurrence of the named job with level Full. If you have a single job definition doing both standard and offsite backups, then there is a danger that the differential will be based relative to the offsite backup. It might be better to define a separate job for the offsite backup, using the same client and fileset as the regular job. __Martin -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users