Re: [Bacula-users] TR: Getting auto-label feature to work properly
On 12/04/10 19:19, Dominique Jeannerod wrote: One up on this question please. I can’t get the auto-labeling function to work as expected Works for me on bacula 3.0.x. Does anyone use the same functions : backup to disk, auto-label of volumes, and auto-recycling ? Yes. What happens now is that old volumes are recycled as they should be, but new volumes are not created when needed, and I’m getting the famous message : You've probably just set retention periods in your pools such that bacula isn't allowed to recycle any of the used volumes, nor is it allowed to create any new ones (due to max volumes) so it's asking you for help. Try list volumes and look at the retention periods. Remember that at any time you must have one volume free and ready for recycling. If you have retention of 4 weeks, a schedule that runs one job a week, and max volumes = 4 in the pool you *will* run out of volumes, because one hasn't been purged before the next run. Set your retention to 22 days (3*7+1) or set max volumes to 5. Another cause for running out of volumes is when you've manually run a job which has a pool with quite tight retention periods and max volumes settings. The manually run job uses a volume from the pool, so later there aren't enough free so everything grinds to a halt. Hard to say what Bacula *should* do in this situation, but it's current approach ain't helpful. I really wish bacula, when it ran out of volumes, would emit a useful message like: Bacula-dir: Unable to obtain volume from pool {pool-name}. No existing volumes may be recycled due to retention rules, and max volumes = {value} prevents automatic labeling of a new volume. Instead, it just emails you to ask you to mount a volume, which isn't exactly helpful. -- Craig ringer -- Download Intel#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] TR: Getting auto-label feature to work properly
On 13/04/10 16:29, Dominique Jeannerod wrote: Thanks a lot Craig, for your answer. I didn't configure a maximum volumes directive, so what I was expecting Bacula to do is : - Auto-Recycle a volume if possible, regarding retention periods. - Automatically create a new one with auto-label feature I there an implicit maximum volumes limit ? You know, I don't think I've ever tried using a pool without an explicit max volumes. As far as I know there's no built-in limit. The documentation: http://www.bacula.org/fr/dev-manual/Automatic_Volume_Recycling.html suggests that your pool definitions: Pool { Name = PS_Default Pool Type = Backup Recycle = yes AutoPrune = yes VolumeRetention = 3 days Use Volume Once = yes Label Format = SYS-SD-FR-2-${C_SYS+} NextPool = PS_VFull } Pool { Name = PS_VFull Pool Type = Backup Recycle = yes AutoPrune = yes VolumeRetention = 3 days Use Volume Once = yes Label Format = SYS-VFULL-SD-FR-2-${C_SYSVFULL+} Storage = fr-201-sd-vfull } ... should result in proper recycling. Is the problem with a migration job? Or a regular backup job? What does list volumes show? ( This sort of thing is why I *REALLY* wish Bacula would say *why* it wanted a volume mount when it's using an auto-label pool. It's like going next door and saying er, can I borrow a bucket? when what you want to be saying is help, help, my house is on fire! ) -- Craig Ringer -- Download Intel#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] TR: Getting auto-label feature to work properly
Thanks a lot Craig, for your answer. I didn't configure a maximum volumes directive, so what I was expecting Bacula to do is : - Auto-Recycle a volume if possible, regarding retention periods. - Automatically create a new one with auto-label feature I there an implicit maximum volumes limit ? -Message d'origine- De : Craig Ringer [mailto:cr...@postnewspapers.com.au] Envoyé : mardi 13 avril 2010 10:24 À : Dominique Jeannerod Cc : bacula-users@lists.sourceforge.net Objet : Re: [Bacula-users] TR: Getting auto-label feature to work properly On 12/04/10 19:19, Dominique Jeannerod wrote: One up on this question please. I cant get the auto-labeling function to work as expected Works for me on bacula 3.0.x. Does anyone use the same functions : backup to disk, auto-label of volumes, and auto-recycling ? Yes. What happens now is that old volumes are recycled as they should be, but new volumes are not created when needed, and Im getting the famous message : You've probably just set retention periods in your pools such that bacula isn't allowed to recycle any of the used volumes, nor is it allowed to create any new ones (due to max volumes) so it's asking you for help. Try list volumes and look at the retention periods. Remember that at any time you must have one volume free and ready for recycling. If you have retention of 4 weeks, a schedule that runs one job a week, and max volumes = 4 in the pool you *will* run out of volumes, because one hasn't been purged before the next run. Set your retention to 22 days (3*7+1) or set max volumes to 5. Another cause for running out of volumes is when you've manually run a job which has a pool with quite tight retention periods and max volumes settings. The manually run job uses a volume from the pool, so later there aren't enough free so everything grinds to a halt. Hard to say what Bacula *should* do in this situation, but it's current approach ain't helpful. I really wish bacula, when it ran out of volumes, would emit a useful message like: Bacula-dir: Unable to obtain volume from pool {pool-name}. No existing volumes may be recycled due to retention rules, and max volumes = {value} prevents automatic labeling of a new volume. Instead, it just emails you to ask you to mount a volume, which isn't exactly helpful. -- Craig ringer -- Download Intel#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] TR: Getting auto-label feature to work properly
Yes, I also understood with the documentation that my configuration should work. I have the same problem with 3.0.3, and 5.0.1 versions. The problem occurs with regular backup jobs, only when scheduled (or ran in parallel), as manual tests on a single job are ok (volumes are being recycled, and then new volumes auto-labeled). -Message d'origine- De : Craig Ringer [mailto:cr...@postnewspapers.com.au] Envoyé : mardi 13 avril 2010 10:47 À : Dominique Jeannerod Cc : bacula-users@lists.sourceforge.net Objet : Re: [Bacula-users] TR: Getting auto-label feature to work properly On 13/04/10 16:29, Dominique Jeannerod wrote: Thanks a lot Craig, for your answer. I didn't configure a maximum volumes directive, so what I was expecting Bacula to do is : - Auto-Recycle a volume if possible, regarding retention periods. - Automatically create a new one with auto-label feature I there an implicit maximum volumes limit ? You know, I don't think I've ever tried using a pool without an explicit max volumes. As far as I know there's no built-in limit. The documentation: http://www.bacula.org/fr/dev-manual/Automatic_Volume_Recycling.html suggests that your pool definitions: Pool { Name = PS_Default Pool Type = Backup Recycle = yes AutoPrune = yes VolumeRetention = 3 days Use Volume Once = yes Label Format = SYS-SD-FR-2-${C_SYS+} NextPool = PS_VFull } Pool { Name = PS_VFull Pool Type = Backup Recycle = yes AutoPrune = yes VolumeRetention = 3 days Use Volume Once = yes Label Format = SYS-VFULL-SD-FR-2-${C_SYSVFULL+} Storage = fr-201-sd-vfull } ... should result in proper recycling. Is the problem with a migration job? Or a regular backup job? What does list volumes show? ( This sort of thing is why I *REALLY* wish Bacula would say *why* it wanted a volume mount when it's using an auto-label pool. It's like going next door and saying er, can I borrow a bucket? when what you want to be saying is help, help, my house is on fire! ) -- Craig Ringer -- Download Intel#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] TR: Getting auto-label feature to work properly
Try using 'LabelFormat' without space? On 12/04/10 19:19, Dominique Jeannerod wrote: One up on this question please. I can’t get the auto-labeling function to work as expected -- Download Intel#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] TR: Getting auto-label feature to work properly
One up on this question please. I cant get the auto-labeling function to work as expected Does anyone use the same functions : backup to disk, auto-label of volumes, and auto-recycling ? It seems to me quite a standard way of using Bacula. What am I doing wrong ? De : Dominique Jeannerod [mailto:dominique.jeanne...@interact-iv.com] Envoyé : lundi 22 mars 2010 18:12 À : 'bacula-users@lists.sourceforge.net' Objet : Getting auto-label feature to work properly Hello, Im currently working on Bacula 5.0.1 to upgrade our current production backup system. We are backuping on disk files, and thus implement the auto-label feature, using counters. On our first tests, everything was running as planned : volumes were created and labeled automatically, and then recycled. What happens now is that old volumes are recycled as they should be, but new volumes are not created when needed, and Im getting the famous message : Please use the label command to create a new Volume for: Storage: fr-201-sd-2 (/backups) Pool: PS_Default Media type: File I really dont understand why ? Our version : Bacula 3.0.3 on CentOS 5.4 Compilated from source with all pre-requisites: yum install httpd mysql-server mysql-devel php mtx gcc gcc-c++ readline-devel make openssl-devel cd /root/bacula-3.0.3 make distclean ./configure --sbindir=/opt/bacula/bin --sysconfdir=/opt/bacula/etc --with-pid-dir=/opt/bacula/working --with-subsys-dir=/opt/bacula/working --enable-smartalloc --with-mysql --with-working-dir=/opt/bacula/working --with-dump-email=xx...@x.fr --with-job-email=xx...@x.fr with-smtp-host=localhost make make install make install-autostart Director Configuration Director {# define myself Name = sv-ics-tst-bkp-fr-201-dir DIRport = 9101# where we listen for UA connections QueryFile = /opt/bacula/opt/etc/query.sql WorkingDirectory = /opt/bacula/working PidDirectory = /opt/bacula/working Maximum Concurrent Jobs = 60 Password = XX # Console password Messages = Daemon FD Connect Timeout = 5 minutes# Default : 30 minutes SD Connect Timeout = 5 minutes# Default : 30 minutes } Storage { Name = fr-201-sd-1 Address = 10.3.1.254 SDPort = 9103 Maximum Concurrent Jobs = 1 Password = XX Device = fr-201-sd-1 Media Type = File } Storage { Name = fr-201-sd-2 Address = 10.3.1.254 SDPort = 9103 Maximum Concurrent Jobs = 1 Password = XX Device = fr-201-sd-2 Media Type = File } JobDefs { Name = JS_Default Type = Backup Level = Incremental Client = sv-ics-tst-bkp-fr-201 FileSet = FS_Default Schedule = SS_Default Storage = fr-201-sd-1 Messages = Standard Pool = PS_Default Accurate = Yes Max Wait Time = 14400 # 4 hours Max Run Sched Time = 27000 } FileSet { Name = FS_Default Include { Options { signature = MD5 onefs=no # on ne reste pas sur un seul file-system fstype=ext2 # pour ext3 : fstype=ext2 aussi ... compression = GZIP } File = / } Exclude { File = /tmp File = /.journal File = /.fsck File = /opt } } Schedule { Name = SS_PRD_1 Run = Level=Full Pool=PS_Default thu at 23:45 Run = Level=Full Pool=PS_Default fri-wed at 23:45 } Pool { Name = PS_Default Pool Type = Backup Recycle = yes AutoPrune = yes VolumeRetention = 3 days Use Volume Once = yes Label Format = SYS-SD-FR-2-${C_SYS+} NextPool = PS_VFull } Pool { Name = PS_VFull Pool Type = Backup Recycle = yes AutoPrune = yes VolumeRetention = 3 days Use Volume Once = yes Label Format = SYS-VFULL-SD-FR-2-${C_SYSVFULL+} Storage = fr-201-sd-vfull } Counter { Name = C_SYS Catalog = MyCatalog } Counter { Name = C_SYSVFULL Catalog = MyCatalog } #CLIENTS : All clients have the same conf Client { Name = sv-ics-tst-bkp-fr-201 Address = 10.3.1.254 FDPort = 9102 Catalog = MyCatalog Maximum Concurrent Jobs = 2 Password = XX # password for FileDaemon File Retention = 3 months Job Retention = 3 months AutoPrune = yes # Prune expired Jobs/Files } # JOBS : All Jobs have the same conf except, using different storages (fr-201-sd-1, fr-201-sd-2, ) Job { Name = JS_sv-ics-tst-bkp-fr-201 Enabled = yes JobDefs = JS_Default Client = sv-ics-tst-bkp-fr-201 Write Bootstrap = /opt/bacula/bootstrap/sv-ics-tst-bkp-fr-201.sys.bsr Storage = fr-201-sd-1 Schedule = SS_PRD_1 } STORAGE CONFIGURATION (SD) Storage {