[Bacula-users] Bacula 5.2.4 Major problems with Warnings jobs not added to catalog etc
Dear All Have anyone had similar problems with 5.2.4 with backup warnings - the MAJOR problem seems that the Backup is NOT added to the catalog and as I backup clients to disc and migrate full backups to tape I suspect these jobs will not be migrated - the jobs migration jobs run tonight. I have reverted to 5.2.3 and reran the Full backups of those clients with warnings. I accept there will be dodgy files on Windows clients but would prefer the backup to treated as a valid backup with respect to adding it to the Catalog and as a job valid for migration. When backing up windows clients I have got the following that resulted in a warning - I have copied the backup status from the Volume for 5.2.4 and 5.2.3 --- Bacula: Backup OK -- with warnings of XXX Full - VSS problem Generate VSS snapshot of drive d:\ failed. VSS support is disabled on this drive. With 5.2.4 238,776 | civengXXX | 2012-01-23 10:05:00 | B| F |302 | 1,793,433,275 | W With 5.2.3 238,776 | civengXXX | 2012-01-23 10:05:00 | B| F |302 | 1,793,433,275 | W Bacula: Backup OK -- with warnings of YYY Full - encrypted file Cannot open D:/Data/Sattar/others/Programming/matlab/yyy.m: ERR=Access is denied. With 5.2.4 238,790 | civengYYY | 2012-01-23 10:05:02 | B| F | 11,465 | 2,255,073,886 | W With 5.2.3 238,799 | civengYYY | 2012-01-23 12:11:05 | B| F | 11,465 | 2,255,185,607 | T --- Bacula: Backup Unknown term code of civengAAA Full - OSX folder D:/Data/My Documents/WDSA 2012/Notices, Banner/full-colour-vert/__MACOSX/._UoA_col_vert.png: ERR=Access is denied. With 5.2.4 238,771 | civengAAA| 2012-01-23 10:00:09 | B| F | 42,781 | 46,811,644,364 | W With 5.2.3 | 238,797 | civengAAA| 2012-01-23 11:51:54 | B| F | 42,782 | 46,813,945,624 | T Regards Stephen Carr -- Stephen Carr Computing Officer School of Civil and Environmental Engineering The University of Adelaide Tel +618-8303-4313 Fax +618-8303-4359 Email sgc...@civeng.adelaide.edu.au CRICOS Provider Number 00123M --- This email message is intended only for the addressee(s)and contains information that may be confidential and/or copyright. If you are not the intended recipient please notify the sender by reply email and immediately delete this email. Use, disclosure or reproduction of this email by anyone other than the intended recipient(s) is strictly prohibited. No representation is made that this email or any attachments are free of viruses. Virus scanning is recommended and is the responsibility of the recipient. -- Try before you buy = See our experts in action! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-dev2___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Fedora 16 - RHEL 5/6 Bacula RPM repository
Hello, please have a look at the readme file at: http://repos.fedorapeople.org/repos/slaanesh/bacula/README.txt there's this note: ** The included /usr/share/doc/bacula-common-%{version}/README.Fedora contains quick installation instructions and notes ** You'll find your quick answer by reading it. Regards, --Simone On 23 January 2012 08:24, tonyalbers bacula-fo...@backupcentral.com wrote: Hi all, Simone, are the RHEL 6 packages compiled with mysql support? Whenever I try to start the director, i get this message in the log file: 22-Jan 17:43 bacula-dir JobId 0: Fatal error: postgresql.c:241 Unable to connect to PostgreSQL server. Database=bacula User=bacula Possible causes: SQL server not running; password incorrect; max_connections exceeded. But I'm running mysql, and it is working as it should be. Should I specify which database server I want to use somewhere? /tony +-- |This was sent by tony.alb...@gmail.com via Backup Central. |Forward SPAM to ab...@backupcentral.com. +-- -- Try before you buy = See our experts in action! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-dev2 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- You cannot discover new oceans unless you have the courage to lose sight of the shore (R. W. Emerson). -- Try before you buy = See our experts in action! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-dev2 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Bacula 5.2.4 problem with warnings
Dear All Has anyone had similar problems with 5.2.4 with backup warnings. The MAJOR problem seems that the Backup is NOT added to the catalog and as I backup clients to disc volume and then migrate Full backups to tape I suspect these jobs will not be migrated - the migration jobs will run tonight. I have just noticed that a job with warnings was migrated. Why is it not in the catalog prior to migration? What if the job is not Migrated - applies to my Incremental backups that remain on disc? I have reverted to 5.2.3 and reran the Full backups of those clients with warnings. I accept there will be dodgy files on Windows clients but would prefer the backup to treated as a valid backup with respect to adding it to the Catalog and as a job valid for migration. When backing up windows clients I have got the following that resulted in a warning - I have copied the backup status from the Volume for 5.2.4 and 5.2.3 --- Bacula: Backup OK -- with warnings of XXX Full - VSS problem Generate VSS snapshot of drive d:\ failed. VSS support is disabled on this drive. With 5.2.4 238,776 | civengXXX | 2012-01-23 10:05:00 | B| F |302 | 1,793,433,275 | W With 5.2.3 238,796 | civengXXX | 2012-01-23 11:50:32 | B| F |302 | 1,793,514,892 | T Using Query option 5 for civengXXX after rerun Note JobID 238,776 is missing | 238,337 | civengXXX | workstation | I | 2012-01-20 14:00:06 | 14 | 566,133,909 | Inc1232| | 238,796 | civengXXX | workstation | F | 2012-01-23 11:50:32 | 302 | 1,793,514,892 | Full1135 | Bacula: Backup OK -- with warnings of YYY Full - encrypted file Cannot open D:/Data/Sattar/others/Programming/matlab/yyy.m: ERR=Access is denied. With 5.2.4 238,790 | civengYYY | 2012-01-23 10:05:02 | B| F | 11,465 | 2,255,073,886 | W With 5.2.3 238,799 | civengYYY | 2012-01-23 12:11:05 | B| F | 11,465 | 2,255,185,607 | T --- Bacula: Backup Unknown term code of civengAAA Full - OSX folder D:/Data/My Documents/WDSA 2012/Notices, Banner/full-colour-vert/__MACOSX/._UoA_col_vert.png: ERR=Access is denied. With 5.2.4 238,771 | civengAAA| 2012-01-23 10:00:09 | B| F | 42,781 | 46,811,644,364 | W With 5.2.3 238,797 | civengAAA| 2012-01-23 11:51:54 | B| F | 42,782 | 46,813,945,624 | T This is the client that has had both the backup with warnings and a normal (5.2.3) backup done Using Query option 5 for civengAAA after migration | 233,051 | civengAAA | workstation | I | 2011-12-21 10:02:09 | 136 | 3,457,909,805 | Inc0011| | 238,895 | civengAAA | workstation | F | 2012-01-23 10:00:09 | 42,781 | 46,821,266,974 | LTO-D42| | 238,890 | civengAAA | workstation | F | 2012-01-23 11:51:54 | 42,782 | 46,823,568,423 | LTO-D42| Regards Stephen Carr -- Stephen Carr Computing Officer School of Civil and Environmental Engineering The University of Adelaide Tel +618-8303-4313 Fax +618-8303-4359 Email sgc...@civeng.adelaide.edu.au CRICOS Provider Number 00123M --- This email message is intended only for the addressee(s)and contains information that may be confidential and/or copyright. If you are not the intended recipient please notify the sender by reply email and immediately delete this email. Use, disclosure or reproduction of this email by anyone other than the intended recipient(s) is strictly prohibited. No representation is made that this email or any attachments are free of viruses. Virus scanning is recommended and is the responsibility of the recipient. -- Stephen Carr Computing Officer School of Civil and Environmental Engineering The University of Adelaide Tel +618-8303-4313 Fax +618-8303-4359 Email sgc...@civeng.adelaide.edu.au CRICOS Provider Number 00123M --- This email message is intended only for the addressee(s)and contains information that may be confidential and/or copyright. If you are not the intended recipient please notify the sender by reply email and immediately delete this email. Use, disclosure or reproduction of this email by anyone other than the intended recipient(s) is strictly prohibited. No representation is made that this email or any attachments are free of viruses. Virus scanning is recommended and is the responsibility of the recipient. -- Try before you buy = See our experts in action! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-dev2___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Bacula 5.2.4 Major problems with Warnings jobs not added to catalog etc
Hello Stephen, On 23/01/2012 09:08, Stephen G Carr wrote: Dear All Have anyone had similar problems with 5.2.4 with backup warnings - the MAJOR problem seems that the Backup is NOT added to the catalog and as I backup clients to disc and migrate full backups to tape I suspect these jobs will not be migrated - the jobs migration jobs run tonight. I would be nice to not just suspect when you are reporting a potential problem such as this one. Did you try to migrate new jobs? Did you get any problem related to Bacula? I have reverted to 5.2.3 and reran the Full backups of those clients with warnings. I accept there will be dodgy files on Windows clients but would prefer the backup to treated as a valid backup with respect to adding it to the Catalog and as a job valid for migration. The Warning job status is treated like the Ok status. Unless you are writing your own queries to migrate jobs, you shouldn't have problems. So, if you are using a custom query, you can just replace JobStatus = 'T' by JobStatus IN ('T', 'W') as we did in other part of Bacula. Bye -- Need professional help and support for Bacula ? Visit http://www.baculasystems.com -- Try before you buy = See our experts in action! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-dev2 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Fedora 16 - RHEL 5/6 Bacula RPM repository
please have a look at the readme file at: http://repos.fedorapeople.org/repos/slaanesh/bacula/README.txt there's this note: ** The included /usr/share/doc/bacula-common-%{version}/README.Fedora contains quick installation instructions and notes ** You'll find your quick answer by reading it. Thanks Simone, sorry for not reading the readme file. I thought it was only for Fedora. /tony +-- |This was sent by tony.alb...@gmail.com via Backup Central. |Forward SPAM to ab...@backupcentral.com. +-- -- Try before you buy = See our experts in action! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-dev2 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Bacula 5.2.4 Major problems with Warnings jobs not added to catalog etc
Dear Eric Thanks for the clue regarding queries. I have just looked at the default set of queries I use (examples/sample-query.sql) and noticed statements like # 5 :List all backups for a Client *Enter Client Name: SELECT DISTINCT Job.JobId as JobId,Client.Name as Client, FileSet.FileSet AS FileSet,Level,StartTime, JobFiles,JobBytes,VolumeName FROM Client,Job,JobMedia,Media,FileSet WHERE Client.Name='%1' AND Client.ClientId=Job.ClientId AND Job.Type='B' AND Job.JobStatus='T' AND Job.FileSetId=FileSet.FileSetId AND JobMedia.JobId=Job.JobId AND JobMedia.MediaId=Media.MediaId ORDER BY Job.StartTime; Note the Job.JobStatus='T' - need to change to (T,W) as you said. I retract the that there is a problem with Bacula and warnings - it is the query !! Sorry Stephen Carr Eric Bollengier wrote: Hello, On 23/01/2012 11:37, Stephen G Carr wrote: Dear Eric I sent a later email and for got to add Update to the subject - the migration of jobs with warnings do migrate to tape. I am using the standard queries - will look at the syntax but why is a job shown in a Volume with warnings but not visible when querying the client? Sorry, I don't understand where jobs are not visible, we might have a display problem somewhere like in Bat or Bweb, but the core code should handle everything properly. Bye Thanks Stephen Carr Eric Bollengier wrote: Hello Stephen, On 23/01/2012 09:08, Stephen G Carr wrote: Dear All Have anyone had similar problems with 5.2.4 with backup warnings - the MAJOR problem seems that the Backup is NOT added to the catalog and as I backup clients to disc and migrate full backups to tape I suspect these jobs will not be migrated - the jobs migration jobs run tonight. I would be nice to not just suspect when you are reporting a potential problem such as this one. Did you try to migrate new jobs? Did you get any problem related to Bacula? I have reverted to 5.2.3 and reran the Full backups of those clients with warnings. I accept there will be dodgy files on Windows clients but would prefer the backup to treated as a valid backup with respect to adding it to the Catalog and as a job valid for migration. The Warning job status is treated like the Ok status. Unless you are writing your own queries to migrate jobs, you shouldn't have problems. So, if you are using a custom query, you can just replace JobStatus = 'T' by JobStatus IN ('T', 'W') as we did in other part of Bacula. Bye -- Stephen Carr Computing Officer School of Civil and Environmental Engineering The University of Adelaide Tel +618-8303-4313 Fax +618-8303-4359 emailsgc...@civeng.adelaide.edu.au CRICOS Provider Number 00123M --- This email message is intended only for the addressee(s)and contains information that may be confidential and/or copyright. If you are not the intended recipient please notify the sender by reply email and immediately delete this email. Use, disclosure or reproduction of this email by anyone other than the intended recipient(s) is strictly prohibited. No representation is made that this email or any attachments are free of viruses. Virus scanning is recommended and is the responsibility of the recipient. -- Stephen Carr Computing Officer School of Civil and Environmental Engineering The University of Adelaide Tel +618-8303-4313 Fax +618-8303-4359 Email sgc...@civeng.adelaide.edu.au CRICOS Provider Number 00123M --- This email message is intended only for the addressee(s)and contains information that may be confidential and/or copyright. If you are not the intended recipient please notify the sender by reply email and immediately delete this email. Use, disclosure or reproduction of this email by anyone other than the intended recipient(s) is strictly prohibited. No representation is made that this email or any attachments are free of viruses. Virus scanning is recommended and is the responsibility of the recipient. -- Try before you buy = See our experts in action! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-dev2 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Have we reached bacula's limits?
Hi folks, we're running four bacula installations, most of them on version 5.0.x compiled from source on CentOS 5.x / 6.x 64bit servers. We're mostly happy with the setup, backups are fast, reliable and generally do not cause us a lot of headaches. Today, a colleague asked me to restore some data from the last backup of a client on our largest bacula install, namely (according to bweb) DB Size: Total clients: 107 Total bytes stored: 34.41 TB Total files:47495362 Database size:31.64 GB MySQL isn't exactly huge, and restoring the data didn't look like too much of a big deal at first: +---+---+--++-+--+ | 9,582 | F | 527,265 | 55,999,595,573 | 2012-01-20 21:05:03 | OFFLINE14_02 | | 9,652 | I |1,150 | 1,534,499,185 | 2012-01-21 18:34:56 | OFFLINE15_01 | +---+---+--++-+--+ So we're talking a mere 500,000 files (he only needed a single dir out of the bunch). 4,5 hours later, Bacula is still sitting at the Building Directory Tree message, without so much as a single . or + hopefully showing up in the terminal, indicating some kind of progress. I've run mysqltuner on the db a couple of times as this isn't the first time we've had problems during a restore, and it looks ok (to my untrained, non-dba eyes anyway): ## General Statistics -- [--] Skipped version check for MySQLTuner script [OK] Currently running supported MySQL version 5.1.52-log [OK] Operating on 64-bit architecture Storage Engine Statistics --- [--] Status: -Archive -BDB -Federated -InnoDB -ISAM -NDBCluster [--] Data in MyISAM tables: 31G (Tables: 33) [!!] Total fragmented tables: 2 Performance Metrics - [--] Up for: 35s (57 q [1.629 qps], 12 conn, TX: 44K, RX: 3K) [--] Reads / Writes: 100% / 0% [--] Total buffers: 12.0G global + 83.2M per thread (151 max threads) [!!] Maximum possible memory usage: 24.3G (137% of installed RAM) [OK] Slow queries: 0% (0/57) [OK] Highest usage of available connections: 0% (1/151) [OK] Key buffer size / total MyISAM indexes: 11.9G/15.3G [OK] Key buffer hit rate: 100.0% (6K cached / 2 reads) [!!] Query cache efficiency: 0.0% (0 cached / 23 selects) [OK] Query cache prunes per day: 0 [OK] Sorts requiring temporary tables: 0% (0 temp sorts / 9 sorts) [!!] Temporary tables created on disk: 34% (8 on disk / 23 total) [OK] Thread cache hit rate: 91% (1 created / 12 connections) [OK] Table cache hit rate: 85% (41 open / 48 opened) [OK] Open file limit used: 1% (83/4K) [OK] Table locks acquired immediately: 100% (38 immediate / 38 locks) [!!] Connections aborted: 8% Recommendations - General recommendations: Run OPTIMIZE TABLE to defragment tables for better performance MySQL started within last 24 hours - recommendations may be inaccurate Reduce your overall MySQL memory footprint for system stability When making adjustments, make tmp_table_size/max_heap_table_size equal Reduce your SELECT DISTINCT queries without LIMIT clauses Your applications are not closing MySQL connections properly Variables to adjust: *** MySQL's maximum memory usage is dangerously high *** *** Add RAM before increasing MySQL buffer variables *** query_cache_limit ( 16M, or use smaller result sets) tmp_table_size ( 61M) max_heap_table_size ( 16M) ## For the restore run mentioned above, I'm seeing a 40MB mysql tmp table in /tmp updated every once in a while, and there's lots of write activity to the partition that holds /tmp. I'm now running a repair table File after cancelling the restore job, but I guess there's something seriously wrong with the above setup. The other bacula servers generally run on smaller machines, but come up with a dir tree after five to ten minutes for a comparable job which is acceptable, but 5 hours seems way off the mark. the bacula db was created using bacula's own mysql init script, so I assume all the indices where created (and more importantly, no extra ones that might slow bacula down) correctly. Insert performance it great during backups, we usually achieve around 30-50MB/sec sustained for 3 to 4 jobs running in parallel. Mem usage (as opposed to mysql tuner's warning) is ok during the restore run, no swapping, 5GB of 18G total still used for buffer cache. Thanks in advance for any help / hints or thoughts, Uwe PS: Please let me know if I should provide more info on the setup what would help in analyzing this problem. -- NIONEX --- Ein Unternehmen der Bertelsmann AG
Re: [Bacula-users] Have we reached bacula's limits?
On 01/23/2012 10:28 AM, Uwe Schuerkamp wrote: I've run mysqltuner on the db a couple of times as this isn't the first time we've had problems during a restore, and it looks ok (to my untrained, non-dba eyes anyway): ## General Statistics -- [--] Skipped version check for MySQLTuner script [OK] Currently running supported MySQL version 5.1.52-log [OK] Operating on 64-bit architecture Storage Engine Statistics --- [--] Status: -Archive -BDB -Federated -InnoDB -ISAM -NDBCluster [--] Data in MyISAM tables: 31G (Tables: 33) [!!] Total fragmented tables: 2 Performance Metrics - [--] Up for: 35s (57 q [1.629 qps], 12 conn, TX: 44K, RX: 3K) [--] Reads / Writes: 100% / 0% [--] Total buffers: 12.0G global + 83.2M per thread (151 max threads) [!!] Maximum possible memory usage: 24.3G (137% of installed RAM) [OK] Slow queries: 0% (0/57) [OK] Highest usage of available connections: 0% (1/151) [OK] Key buffer size / total MyISAM indexes: 11.9G/15.3G [OK] Key buffer hit rate: 100.0% (6K cached / 2 reads) [!!] Query cache efficiency: 0.0% (0 cached / 23 selects) [OK] Query cache prunes per day: 0 [OK] Sorts requiring temporary tables: 0% (0 temp sorts / 9 sorts) [!!] Temporary tables created on disk: 34% (8 on disk / 23 total) [OK] Thread cache hit rate: 91% (1 created / 12 connections) [OK] Table cache hit rate: 85% (41 open / 48 opened) [OK] Open file limit used: 1% (83/4K) [OK] Table locks acquired immediately: 100% (38 immediate / 38 locks) [!!] Connections aborted: 8% Recommendations - General recommendations: Run OPTIMIZE TABLE to defragment tables for better performance MySQL started within last 24 hours - recommendations may be inaccurate Reduce your overall MySQL memory footprint for system stability When making adjustments, make tmp_table_size/max_heap_table_size equal Reduce your SELECT DISTINCT queries without LIMIT clauses Your applications are not closing MySQL connections properly Variables to adjust: *** MySQL's maximum memory usage is dangerously high *** *** Add RAM before increasing MySQL buffer variables *** query_cache_limit ( 16M, or use smaller result sets) tmp_table_size ( 61M) max_heap_table_size ( 16M) ## For the restore run mentioned above, I'm seeing a 40MB mysql tmp table in /tmp updated every once in a while, and there's lots of write activity to the partition that holds /tmp. If max_heap_table_size is 16M, then in-memory temporary tables are limited to 16M too. Maximum in-memory temporary table size is the smaller of tmp_table-size and max_heap_table_size. You only ever have a single DB connection; why are you allowing 151 connections? Cut max_connections to 10, increase tmp_table_size and max_heap_table_size to 64M or even 128M, increase table_cache to 64, disable the query cache because you're going to have few if any frequently-repeated queries, update to MySQL 5.5, and seriously, seriously consider converting to InnoDB. It is a MUCH higher performance storage engine than MyISAM. Remember that MyISAM was designed to yield *acceptable* performance in shared installations on machines with less than 32MB of RAM. -- Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355 ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org Renaissance Man, Unix ronin, Perl hacker, SQL wrangler, Free Stater It's not the years, it's the mileage. -- Try before you buy = See our experts in action! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-dev2 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Have we reached bacula's limits?
1: Make sure you have enough ram in your mysql box (ie, several 10s of Gb) 2: Make sure you tune mysql properly. Most of the supplied config examples are for sub-1Gb memory configuration. 3: Make sure you have the _correct_ indexes built. this is in the bacula knowledgebase. 4: For systems with 10s of millions of files - Seriously consider moving to postgres. MySQL is a memory hog. On 23/01/12 15:28, Uwe Schuerkamp wrote: Hi folks, we're running four bacula installations, most of them on version 5.0.x compiled from source on CentOS 5.x / 6.x 64bit servers. We're mostly happy with the setup, backups are fast, reliable and generally do not cause us a lot of headaches. Today, a colleague asked me to restore some data from the last backup of a client on our largest bacula install, namely (according to bweb) DB Size: Total clients:107 Total bytes stored: 34.41 TB Total files: 47495362 Database size:31.64 GB MySQL isn't exactly huge, and restoring the data didn't look like too much of a big deal at first: +---+---+--++-+--+ | 9,582 | F | 527,265 | 55,999,595,573 | 2012-01-20 21:05:03 | OFFLINE14_02 | | 9,652 | I |1,150 | 1,534,499,185 | 2012-01-21 18:34:56 | OFFLINE15_01 | +---+---+--++-+--+ So we're talking a mere 500,000 files (he only needed a single dir out of the bunch). 4,5 hours later, Bacula is still sitting at the Building Directory Tree message, without so much as a single . or + hopefully showing up in the terminal, indicating some kind of progress. I've run mysqltuner on the db a couple of times as this isn't the first time we've had problems during a restore, and it looks ok (to my untrained, non-dba eyes anyway): ## General Statistics -- [--] Skipped version check for MySQLTuner script [OK] Currently running supported MySQL version 5.1.52-log [OK] Operating on 64-bit architecture Storage Engine Statistics --- [--] Status: -Archive -BDB -Federated -InnoDB -ISAM -NDBCluster [--] Data in MyISAM tables: 31G (Tables: 33) [!!] Total fragmented tables: 2 Performance Metrics - [--] Up for: 35s (57 q [1.629 qps], 12 conn, TX: 44K, RX: 3K) [--] Reads / Writes: 100% / 0% [--] Total buffers: 12.0G global + 83.2M per thread (151 max threads) [!!] Maximum possible memory usage: 24.3G (137% of installed RAM) [OK] Slow queries: 0% (0/57) [OK] Highest usage of available connections: 0% (1/151) [OK] Key buffer size / total MyISAM indexes: 11.9G/15.3G [OK] Key buffer hit rate: 100.0% (6K cached / 2 reads) [!!] Query cache efficiency: 0.0% (0 cached / 23 selects) [OK] Query cache prunes per day: 0 [OK] Sorts requiring temporary tables: 0% (0 temp sorts / 9 sorts) [!!] Temporary tables created on disk: 34% (8 on disk / 23 total) [OK] Thread cache hit rate: 91% (1 created / 12 connections) [OK] Table cache hit rate: 85% (41 open / 48 opened) [OK] Open file limit used: 1% (83/4K) [OK] Table locks acquired immediately: 100% (38 immediate / 38 locks) [!!] Connections aborted: 8% Recommendations - General recommendations: Run OPTIMIZE TABLE to defragment tables for better performance MySQL started within last 24 hours - recommendations may be inaccurate Reduce your overall MySQL memory footprint for system stability When making adjustments, make tmp_table_size/max_heap_table_size equal Reduce your SELECT DISTINCT queries without LIMIT clauses Your applications are not closing MySQL connections properly Variables to adjust: *** MySQL's maximum memory usage is dangerously high *** *** Add RAM before increasing MySQL buffer variables *** query_cache_limit ( 16M, or use smaller result sets) tmp_table_size ( 61M) max_heap_table_size ( 16M) ## For the restore run mentioned above, I'm seeing a 40MB mysql tmp table in /tmp updated every once in a while, and there's lots of write activity to the partition that holds /tmp. I'm now running a repair table File after cancelling the restore job, but I guess there's something seriously wrong with the above setup. The other bacula servers generally run on smaller machines, but come up with a dir tree after five to ten minutes for a comparable job which is acceptable, but 5 hours seems way off the mark. the bacula db was created using bacula's own mysql init script, so I assume all the indices where created (and more importantly, no extra ones that might slow
Re: [Bacula-users] Will Bacula be faster with Two Tape Devices?
But will this speed up my backups? Generally speaking, should Bacula be able to write two the two tapes devices more quickly than it can write to one? The unit in question is a Tandberg Data StorageLibrary T24 LTO with two serial attached scsi LTO4 tape devices, attached to a server running Centos 6.2. It will not speed up a single job since 1 job can not use more than 1 drive however if you have concurrent jobs with spooling and you can provide enough bandwidth to keep up with the drives it can speed up multiple jobs. Remember that at 2:1 compression these LTO4 drives write at 120MB /s each so you will need to have a raid array (raid 0 maybe) or fast SSD for your spool location and this raid probably should not be on the same drives as your source data. John -- Try before you buy = See our experts in action! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-dev2 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] SCSI Errors
I have Bacula 5.0.3 on FreeBSD 8.2 backing up to an IBM Ultrium ULT3580 tape drive, using 400~800GB tapes, connected to an Adaptec 39320LPE Ultra320 SCSI adapter. I'm using SQLite and have 4GB of RAM. I used to have a Quantum DLTV4 tape drive, but I was getting SCSI errors every few weeks, so I switched to the Ultrium. I've used 3 other tapes in this drive with no issues. This tape fails: # cat /var/log/messages Jan 21 00:44:00 bsd kernel: (sa0:ahd0:0:6:0): WRITE FILEMARKS(6). CDB: 10 0 0 0 1 0 Jan 21 00:44:00 bsd kernel: (sa0:ahd0:0:6:0): CAM status: SCSI Status Error Jan 21 00:44:00 bsd kernel: (sa0:ahd0:0:6:0): SCSI status: Check Condition Jan 21 00:44:00 bsd kernel: (sa0:ahd0:0:6:0): SCSI sense: MEDIUM ERROR asc:31,0 (Medium format corrupted) field replaceable unit: 30 Jan 21 00:44:00 bsd kernel: (sa0:ahd0:0:6:0): WRITE FILEMARKS(6). CDB: 10 0 0 0 2 0 Jan 21 00:44:00 bsd kernel: (sa0:ahd0:0:6:0): CAM status: SCSI Status Error Jan 21 00:44:00 bsd kernel: (sa0:ahd0:0:6:0): SCSI status: Check Condition Jan 21 00:44:00 bsd kernel: (sa0:ahd0:0:6:0): SCSI sense: MEDIUM ERROR asc:31,0 (Medium format corrupted) field replaceable unit: 30 # cat /var/db/bacula/log 21-Jan 00:43 bacula-server-sd JobId 1070: Committing spooled data to Volume BSD2a. Despooling 9,641,289,685 bytes ... 21-Jan 00:43 bacula-server-sd JobId 1070: End of Volume BSD2a at 21:7164 on device ULT3580 (/dev/nsa0). Write of 64512 bytes got 0. 21-Jan 00:44 bacula-server-sd JobId 1070: Error: Error writing final EOF to tape. This Volume may not be readable. dev.c:1745 ioctl MTWEOF error on ULT3580 (/dev/nsa0). ERR=Input/output error. 21-Jan 00:44 bacula-server-sd JobId 1070: End of medium on Volume BSD2a Bytes=13,968,912,384 Blocks=216,531 at 21-Jan-2012 00:44. 21-Jan 00:44 bacula-server-sd JobId 1070: Job nti-bsd.2012-01-20_23.05.00_26 is waiting. Cannot find any appendable volumes. Please use the label command to create a new Volume for: Storage: ULT3580 (/dev/nsa0) Pool: Ultium Media type: LTO4 I've tried migrating the jobs from the tape, but I get: # cat /var/db/bacula/log 23-Jan 15:34 bacula-server-sd JobId 1079: Forward spacing Volume BSD2a to file:block 20:0. 23-Jan 15:35 bacula-server-sd JobId 1079: Error: block.c:1002 Read error on fd=5 at file:blk 21:7163 on device ULT3580 (/dev/nsa0). ERR=Operation not permitted. 23-Jan 15:35 bacula-server-sd JobId 1079: Error: Unexpected Tape is Off-line 23-Jan 15:35 bacula-server-dir JobId 1079: Error: Bacula bacula-server-dir 5.0.3 (04Aug10): 23-Jan-2012 15:35:54 I was able to write multi-tape btape fill to this tape with no errors. I haven't completed the read test, yet. Here's a sample error from my old Quantum DLTV4 drive, since I feel like these might be connected. Dec 14 03:34:35 bsd kernel: (sa0:ahd0:0:5:0): SCSI status: Check Condition Dec 14 03:34:35 bsd kernel: (sa0:ahd0:0:5:0): SCSI sense: UNIT ATTENTION csi:e0,40,0,2e asc:29,3 (Bus device reset function occurred) Dec 14 05:34:46 bsd kernel: ahd0: Recovery Initiated - Card was not paused Other things I've done: * Replaced the SCSI adapter * Replaced the SCSI cable and terminator Here's the current SCSI adapter info # dmesg ahd0: Adaptec 39320LPE Ultra320 SCSI adapter port 0xe400-0xe4ff,0xe800-0xe8ff mem 0xfbd7e000-0xfbd7 irq 16 at device 4.0 on pci4 ahd0: [ITHREAD] aic7902: Ultra320 Wide Channel A, SCSI Id=7, PCI-X 101-133MHz, 512 SCBs Only other thing I haven't done is move to MySQL since my bacula.db is about 3.6 GB holding ~18M files in 332 jobs: select sum(JobFiles) from job; +---+ | sum(JobFiles) | +---+ | 18685071 | +---+ Enter SQL query: select count(*) from job; +--+ | count(*) | +--+ | 332 | +--+ I have no idea where else to look. Any help is appreciated! Nikola Lazic -- Try before you buy = See our experts in action! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-dev2 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] SCSI Errors
On Mon, Jan 23, 2012 at 6:08 PM, Nikola Lazic n...@vpi.us wrote: I have Bacula 5.0.3 on FreeBSD 8.2 backing up to an IBM Ultrium ULT3580 tape drive, using 400~800GB tapes, connected to an Adaptec 39320LPE Ultra320 SCSI adapter. I'm using SQLite and have 4GB of RAM. I used to have a Quantum DLTV4 tape drive, but I was getting SCSI errors every few weeks, so I switched to the Ultrium. I've used 3 other tapes in this drive with no issues. This tape fails: # cat /var/log/messages Jan 21 00:44:00 bsd kernel: (sa0:ahd0:0:6:0): WRITE FILEMARKS(6). CDB: 10 0 0 0 1 0 Jan 21 00:44:00 bsd kernel: (sa0:ahd0:0:6:0): CAM status: SCSI Status Error Jan 21 00:44:00 bsd kernel: (sa0:ahd0:0:6:0): SCSI status: Check Condition Jan 21 00:44:00 bsd kernel: (sa0:ahd0:0:6:0): SCSI sense: MEDIUM ERROR asc:31,0 (Medium format corrupted) field replaceable unit: 30 Jan 21 00:44:00 bsd kernel: (sa0:ahd0:0:6:0): WRITE FILEMARKS(6). CDB: 10 0 0 0 2 0 Jan 21 00:44:00 bsd kernel: (sa0:ahd0:0:6:0): CAM status: SCSI Status Error Jan 21 00:44:00 bsd kernel: (sa0:ahd0:0:6:0): SCSI status: Check Condition Jan 21 00:44:00 bsd kernel: (sa0:ahd0:0:6:0): SCSI sense: MEDIUM ERROR asc:31,0 (Medium format corrupted) field replaceable unit: 30 # cat /var/db/bacula/log 21-Jan 00:43 bacula-server-sd JobId 1070: Committing spooled data to Volume BSD2a. Despooling 9,641,289,685 bytes ... 21-Jan 00:43 bacula-server-sd JobId 1070: End of Volume BSD2a at 21:7164 on device ULT3580 (/dev/nsa0). Write of 64512 bytes got 0. 21-Jan 00:44 bacula-server-sd JobId 1070: Error: Error writing final EOF to tape. This Volume may not be readable. dev.c:1745 ioctl MTWEOF error on ULT3580 (/dev/nsa0). ERR=Input/output error. 21-Jan 00:44 bacula-server-sd JobId 1070: End of medium on Volume BSD2a Bytes=13,968,912,384 Blocks=216,531 at 21-Jan-2012 00:44. 21-Jan 00:44 bacula-server-sd JobId 1070: Job nti-bsd.2012-01-20_23.05.00_26 is waiting. Cannot find any appendable volumes. Please use the label command to create a new Volume for: Storage: ULT3580 (/dev/nsa0) Pool: Ultium Media type: LTO4 I've tried migrating the jobs from the tape, but I get: # cat /var/db/bacula/log 23-Jan 15:34 bacula-server-sd JobId 1079: Forward spacing Volume BSD2a to file:block 20:0. 23-Jan 15:35 bacula-server-sd JobId 1079: Error: block.c:1002 Read error on fd=5 at file:blk 21:7163 on device ULT3580 (/dev/nsa0). ERR=Operation not permitted. 23-Jan 15:35 bacula-server-sd JobId 1079: Error: Unexpected Tape is Off-line 23-Jan 15:35 bacula-server-dir JobId 1079: Error: Bacula bacula-server-dir 5.0.3 (04Aug10): 23-Jan-2012 15:35:54 I was able to write multi-tape btape fill to this tape with no errors. I haven't completed the read test, yet. Here's a sample error from my old Quantum DLTV4 drive, since I feel like these might be connected. Dec 14 03:34:35 bsd kernel: (sa0:ahd0:0:5:0): SCSI status: Check Condition Dec 14 03:34:35 bsd kernel: (sa0:ahd0:0:5:0): SCSI sense: UNIT ATTENTION csi:e0,40,0,2e asc:29,3 (Bus device reset function occurred) Dec 14 05:34:46 bsd kernel: ahd0: Recovery Initiated - Card was not paused Other things I've done: * Replaced the SCSI adapter * Replaced the SCSI cable and terminator Here's the current SCSI adapter info # dmesg ahd0: Adaptec 39320LPE Ultra320 SCSI adapter port 0xe400-0xe4ff,0xe800-0xe8ff mem 0xfbd7e000-0xfbd7 irq 16 at device 4.0 on pci4 ahd0: [ITHREAD] aic7902: Ultra320 Wide Channel A, SCSI Id=7, PCI-X 101-133MHz, 512 SCBs Only other thing I haven't done is move to MySQL since my bacula.db is about 3.6 GB holding ~18M files in 332 jobs: select sum(JobFiles) from job; +---+ | sum(JobFiles) | +---+ | 18685071 | +---+ Enter SQL query: select count(*) from job; +--+ | count(*) | +--+ | 332 | +--+ I have no idea where else to look. Any help is appreciated! Bad tape? Drive need cleaned? John -- Try before you buy = See our experts in action! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-dev2 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] critical error -- tape labels get corrupted, previous backups unreadable
I'm experiencing a critical problem where tape labels on volumes with data get corrupted, leaving all data on the tape inaccessible to bacula. I'm running bacula 5.2.2 built from source, under Linux (CentOS 5.7 x86_64). This problem has happened with approximately 15 tapes over approximately 6 months, mostly new LTO-4 media, but some LTO-3 media that's being reused. The problem is sporadic, appearing in approximately 1 out of 60 tapes per week. I do not think the issue is related to the physical media or the tape drives. One tape was last written successfully when in drive 0, then appears corrupt when a later job tries to use is in drive 1. Another tape was last written successfully when in drive 1, then appears corrupt when a later job tries to use it in drive 0. I'm not sure what combination of circumstances trigger the problem, but it seems to show up after: a tape is used for backups the tape is unloaded the tape is later reloaded for use in another job Data from uncorrupted tapes can be read and restored without problems. Purging the corrupt volume from the bacula database and relabeling media allows me to reuse the tape successfully (at the loss of TB of backups!). A tape will work successfully for many jobs, then later bacula will mount the tape and be unable to read the label. Here are the log records for a particular volume. It was labeled about Dec 22, 2011. First used on Jan 4 2012. Used successfully for 10 jobs (350.49GB), then the label was corrupted. -- 04-Jan 06:24 sbia-infr-vbacula JobId 42676: Using Volume 004090 from 'Scratch' pool. 04-Jan 06:25 sbia-infr-vbacula JobId 42676: Wrote label to prelabeled Volume 004090 on device ml6000-drv1 (/dev/tape1-ml6000) 04-Jan 06:25 sbia-infr-vbacula JobId 42676: New volume 004090 mounted on device ml6000-drv1 (/dev/tape1-ml6000) at 04-Jan-2012 06:25. 04-Jan 08:23 sbia-infr-vbacula JobId 42676: Committing spooled data to Volume 004090. Despooling 37,003,975,390 bytes ... 05-Jan 06:47 sbia-infr-vbacula JobId 42724: Volume 004090 previously written, moving to end of data. 05-Jan 06:49 sbia-infr-vbacula JobId 42724: Ready to append to end of Volume 004090 at file=69. 05-Jan 11:05 sbia-infr-vbacula JobId 42724: Committing spooled data to Volume 004090. Despooling 495 bytes ... 06-Jan 06:51 sbia-infr-vbacula JobId 42746: Volume 004090 previously written, moving to end of data. 06-Jan 06:52 sbia-infr-vbacula JobId 42746: Ready to append to end of Volume 004090 at file=70. 06-Jan 12:08 sbia-infr-vbacula JobId 42746: Committing spooled data to Volume 004090. Despooling 495 bytes ... 07-Jan 06:48 sbia-infr-vbacula JobId 42768: Volume 004090 previously written, moving to end of data. 07-Jan 06:50 sbia-infr-vbacula JobId 42768: Ready to append to end of Volume 004090 at file=71. 07-Jan 12:07 sbia-infr-vbacula JobId 42768: Committing spooled data to Volume 004090. Despooling 495 bytes ... 08-Jan 06:48 sbia-infr-vbacula JobId 42790: Volume 004090 previously written, moving to end of data. 08-Jan 06:50 sbia-infr-vbacula JobId 42790: Ready to append to end of Volume 004090 at file=72. 08-Jan 11:40 sbia-infr-vbacula JobId 42790: Committing spooled data to Volume 004090. Despooling 495 bytes ... 09-Jan 11:12 sbia-infr-vbacula JobId 42812: Committing spooled data to Volume 004090. Despooling 495 bytes ... 10-Jan 05:47 sbia-infr-vbacula JobId 42831: Volume 004090 previously written, moving to end of data. 10-Jan 05:49 sbia-infr-vbacula JobId 42831: Ready to append to end of Volume 004090 at file=74. 10-Jan 23:52 sbia-infr-vbacula JobId 42831: Committing spooled data to Volume 004090. Despooling 35,276,606,182 bytes ... 11-Jan 12:58 sbia-infr-vbacula JobId 42856: Committing spooled data to Volume 004090. Despooling 495 bytes ... 12-Jan 06:45 sbia-infr-vbacula JobId 42879: Volume 004090 previously written, moving to end of data. 12-Jan 06:45 sbia-infr-vbacula JobId 42879: Ready to append to end of Volume 004090 at file=83. 12-Jan 12:56 sbia-infr-vbacula JobId 42879: Committing spooled data to Volume 004090. Despooling 5,914 bytes ... 13-Jan 06:45 sbia-infr-vbacula JobId 42901: Committing spooled data to Volume 004090. Despooling 404 bytes ... 15-Jan 16:54 sbia-infr-vbacula JobId 42924: Please mount Volume 004090 or label a new one for: -- At this point, the volume 004090 is unusable. Running 'btape' on that volume reports [root@sbia-infr1 working]# ../bin/btape -v ml6000-drv0 Tape block granularity is 1024 bytes. btape: butil.c:290 Using device: ml6000-drv0 for writing. 23-Jan 18:14 btape JobId 0: 3301 Issuing autochanger loaded? drive 0 command. 23-Jan 18:14 btape JobId 0: 3302 Autochanger loaded? drive 0, result is Slot 9. btape: btape.c:477 open device ml6000-drv0 (/dev/tape0-ml6000): OK *readlabel btape: btape.c:526 Volume has no label. Volume Label: Id: **error**VerNo : 0 VolName