Re: [Bacula-users] dir and sd hang
On Wednesday 24 June 2009 12:58:21 Attila Fülöp wrote: Silver Salonen wrote: On Tuesday 16 June 2009 15:08:29 Attila Fülöp wrote: Silver Salonen wrote: Hi. I use Bacula 3.0.0 on FreeBSD-6.3. The problem I have is that DIR and SD tend to hang often, and it seems one causes another, because they do it together mostly. Sometimes it happens between jobs, sometimes it happens before all the jobs. The only possibility unlock the processes is to kill them with -9 and start again. After that backups work usually a few days and then the processes hang again. Silver, not sure if you have already seen this: http://security.freebsd.org/advisories/FreeBSD-EN-09:04.fork.asc Maybe this patch will fix your problems with FreeBSD 7.0. Attila I'll try upgrading my 6.2 to 7-STABLE some time. I'll post about the results. -- Silver -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] concurrent backups to one volume?
hi! reading the documentation i understand that you should have several volumes for concurrent backups, on different devices/directories. (i work on disk for now.) However some people here on the list seem to be doing well with concurrent backups to only one volume. is that actually true or am i misunderstanding something? if so, how do i do that? i would like a reciept for that. what bacula version is required for that? /andreas -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Slightly off topic: Using the mailslot on a Quantum Superloader 3.
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hello, Forgive me for the O/T post, but I know several people here are using Quantum Superloaders and have wondered how the mailslot could be used, given that it cannot be addressed via mtx. I found this on the web: http://www.symantec.com/connect/forums/ejecting-tapes-library-after-job-completion-quantum-superloader-3 I removed the 'Windows-ness' and this is what it boils down to: Refreshing the commands page: wget -O /dev/null -q --http-user=guest --http-password=guest http://192.168.1.10/commands.html; Moving tape from Slot 4 to Mailslot: wget -O /dev/null -q --http-user=guest --http-password=guest http://192.168.1.10/move.cgi?from=4to=18; (wait for user to take out tape after doing this) Move tape from Mailslot to Slot 4: wget -O /dev/null -q --http-user=guest --http-password=guest http://192.168.1.10/move.cgi?from=18to=4; (User will have to put tape in and confirm it on the display.) This can be used to get the status page (no auth required) wget -q http://192.168.1.10/status.html; The resulting status.html can be checked for the current Autoloader and drive status (look for Idle to make sure they're not busy before doing any operations .). I plan to use this in combination with 'mtx status' and a few SQL queries to automatically remove Volstatus = 'Full' tapes and insert Recycled ones. (Query SQL to find volume names of 'Full' volumes, cross reference with mtx status to find related slot numbers, eject them to mailslot (one by one), find 'Recycled' volumes to insert, locate free slots and move tape from mailslot there, ...) The whole thing isn't set in stone yet and might fall under the table. If I do complete it, I'll share the resulting scripts here and on the bacula wiki. (Not sure what language I'll be using yet as I'm actually no good at coding.) Greetings, Michel -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (MingW32) iEYEARECAAYFAkpDWEgACgkQ2Vs+MkscAyVc5wCeP3RUT42yVvjP9VgfsxJo6iq0 kfMAoJk2bhlmMTrJPx6s29Us07lWrnCw =sy6D -END PGP SIGNATURE- -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Network send error to SD. ERR=Broken Pipe
Hi guys, does anyone know how to resolve this issue ? Considerations: - Started and restared, all components many times :) - Add hearthbeat interval = 1 to SF and SD - Test connection using telnet, works... - Unix permissions at SD are ok. - This job dies always near this point: ~9min (8.57, 9.03, 9.03) ~11.000 files (11.110, 11.115, 11.113) ~880mb (882, 880, 882) - dir and fd are v1.38 - debian 4 stable - sd is v1.38 on centos - this backup system well worked for 1 year - the architecture: is designed for disaster recovery. there are only 2 server involved: repository and jane. both server have fd and sd, thay are located in 2 different buildings on different subnets. jane-fd sends his backup to repository-sd. repository-fd send his backup to jane-sd. director is located on repository. All advice are welcome. Thanks, Cesare 25-giu 11:28 repository-dir: No prior Full backup Job record found. 25-giu 11:28 repository-dir: No prior or suitable Full backup found. Doing FULL backup. 25-giu 11:28 repository-dir: Start Backup JobId 2248, Job=repository.base.2009-06-25_11.28.07 25-giu 11:09 storage.jane: Volume repository.base-0005 previously written, moving to end of data. 25-giu 11:37 repository-fd: repository.base.2009-06-25_11.28.07 Fatal error: backup.c:500 Network send error to SD. ERR=Pipe rotta 25-giu 11:37 repository-dir: repository.base.2009-06-25_11.28.07 Error: Bacula 1.38.11 (28Jun06): 25-giu-2009 11:37:40 JobId: 2248 Job:repository.base.2009-06-25_11.28.07 Backup Level: Full (upgraded from Incremental) Client: repository i486-pc-linux-gnu,debian,4.0 FileSet:repository.base 2008-03-21 17:35:23 Pool: repository.base Storage:storage.jane Scheduled time: 25-giu-2009 11:28:05 Start time: 25-giu-2009 11:28:43 End time: 25-giu-2009 11:37:40 Elapsed time: 8 mins 57 secs Priority: 10 FD Files Written: 11,113 SD Files Written: 0 FD Bytes Written: 882,597,993 (882.5 MB) SD Bytes Written: 0 (0 B) Rate: 1643,6 KB/s Software Compression: 53,9 % Volume name(s): Volume Session Id: 1 Volume Session Time:1245920820 snap://1245920820 Last Volume Bytes: 1 (1 B) Non-fatal FD errors:0 SD Errors: 0 FD termination status: Error SD termination status: Error Termination:*** Backup Error *** begin:vcard fn:Cesare Montresor n:Montresor;Cesare org:Netspin srl;Research and Development adr:;;Via Maestro Ardizzone, 3;Grezzana;Verona;;Italia email;internet:c.montre...@netspin.it tel;work:045-86 50 147 tel;fax:045-86 58 007 x-mozilla-html:FALSE url:http://www.netspin.it version:2.1 end:vcard -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] concurrent backups to one volume?
On Thursday 25 June 2009 13:42:30 Andreas Schuldei wrote: hi! reading the documentation i understand that you should have several volumes for concurrent backups, on different devices/directories. (i work on disk for now.) However some people here on the list seem to be doing well with concurrent backups to only one volume. is that actually true or am i misunderstanding something? if so, how do i do that? i would like a reciept for that. what bacula version is required for that? /andreas Hello. As one device supports only one job, you have to create separate devices for each job you want to be able to run concurrently. -- Silver -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Network send error to SD. ERR=Broken Pipe
Hi Cesare, I ran accros this issue in 2005, with very similar results. At that stage I was working with Nvida Onboard NICs on the Windows FD. Problem turned out to be driver related. If you are not running Nvidia hardware, investigate and update the driver. Kind Regards Stephan -Original Message- From: Cesare Montresor [mailto:c.montre...@netspin.it] Sent: 25 June 2009 01:46 PM To: Bacula-users@lists.sourceforge.net Subject: [Bacula-users] Network send error to SD. ERR=Broken Pipe Hi guys, does anyone know how to resolve this issue ? Considerations: - Started and restared, all components many times :) - Add hearthbeat interval = 1 to SF and SD - Test connection using telnet, works... - Unix permissions at SD are ok. - This job dies always near this point: ~9min (8.57, 9.03, 9.03) ~11.000 files (11.110, 11.115, 11.113) ~880mb (882, 880, 882) - dir and fd are v1.38 - debian 4 stable - sd is v1.38 on centos - this backup system well worked for 1 year - the architecture: is designed for disaster recovery. there are only 2 server involved: repository and jane. both server have fd and sd, thay are located in 2 different buildings on different subnets. jane-fd sends his backup to repository-sd. repository-fd send his backup to jane-sd. director is located on repository. All advice are welcome. Thanks, Cesare 25-giu 11:28 repository-dir: No prior Full backup Job record found. 25-giu 11:28 repository-dir: No prior or suitable Full backup found. Doing FULL backup. 25-giu 11:28 repository-dir: Start Backup JobId 2248, Job=repository.base.2009-06-25_11.28.07 25-giu 11:09 storage.jane: Volume repository.base-0005 previously written, moving to end of data. 25-giu 11:37 repository-fd: repository.base.2009-06-25_11.28.07 Fatal error: backup.c:500 Network send error to SD. ERR=Pipe rotta 25-giu 11:37 repository-dir: repository.base.2009-06-25_11.28.07 Error: Bacula 1.38.11 (28Jun06): 25-giu-2009 11:37:40 JobId: 2248 Job:repository.base.2009-06-25_11.28.07 Backup Level: Full (upgraded from Incremental) Client: repository i486-pc-linux-gnu,debian,4.0 FileSet:repository.base 2008-03-21 17:35:23 Pool: repository.base Storage:storage.jane Scheduled time: 25-giu-2009 11:28:05 Start time: 25-giu-2009 11:28:43 End time: 25-giu-2009 11:37:40 Elapsed time: 8 mins 57 secs Priority: 10 FD Files Written: 11,113 SD Files Written: 0 FD Bytes Written: 882,597,993 (882.5 MB) SD Bytes Written: 0 (0 B) Rate: 1643,6 KB/s Software Compression: 53,9 % Volume name(s): Volume Session Id: 1 Volume Session Time:1245920820 snap://1245920820 Last Volume Bytes: 1 (1 B) Non-fatal FD errors:0 SD Errors: 0 FD termination status: Error SD termination status: Error Termination:*** Backup Error *** -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Network send error to SD. ERR=Broken Pipe
Cesare Montresor wrote: Hi guys, does anyone know how to resolve this issue ? Considerations: - Started and restared, all components many times :) - Add hearthbeat interval = 1 to SF and SD - Test connection using telnet, works... - Unix permissions at SD are ok. - This job dies always near this point: ~9min (8.57, 9.03, 9.03) ~11.000 files (11.110, 11.115, 11.113) ~880mb (882, 880, 882) - dir and fd are v1.38 - debian 4 stable - sd is v1.38 on centos - this backup system well worked for 1 year If this setup has worked for a year and nothing has been changed recently, then suspect a network hardware problem. - the architecture: is designed for disaster recovery. there are only 2 server involved: repository and jane. both server have fd and sd, thay are located in 2 different buildings on different subnets. jane-fd sends his backup to repository-sd. repository-fd send his backup to jane-sd. director is located on repository. All advice are welcome. Thanks, Cesare 25-giu 11:28 repository-dir: No prior Full backup Job record found. 25-giu 11:28 repository-dir: No prior or suitable Full backup found. Doing FULL backup. 25-giu 11:28 repository-dir: Start Backup JobId 2248, Job=repository.base.2009-06-25_11.28.07 25-giu 11:09 storage.jane: Volume repository.base-0005 previously written, moving to end of data. 25-giu 11:37 repository-fd: repository.base.2009-06-25_11.28.07 Fatal error: backup.c:500 Network send error to SD. ERR=Pipe rotta 25-giu 11:37 repository-dir: repository.base.2009-06-25_11.28.07 Error: Bacula 1.38.11 (28Jun06): 25-giu-2009 11:37:40 JobId: 2248 Job:repository.base.2009-06-25_11.28.07 Backup Level: Full (upgraded from Incremental) Client: repository i486-pc-linux-gnu,debian,4.0 FileSet:repository.base 2008-03-21 17:35:23 Pool: repository.base Storage:storage.jane Scheduled time: 25-giu-2009 11:28:05 Start time: 25-giu-2009 11:28:43 End time: 25-giu-2009 11:37:40 Elapsed time: 8 mins 57 secs Priority: 10 FD Files Written: 11,113 SD Files Written: 0 FD Bytes Written: 882,597,993 (882.5 MB) SD Bytes Written: 0 (0 B) Rate: 1643,6 KB/s Software Compression: 53,9 % Volume name(s): Volume Session Id: 1 Volume Session Time:1245920820 snap://1245920820 Last Volume Bytes: 1 (1 B) Non-fatal FD errors:0 SD Errors: 0 FD termination status: Error SD termination status: Error Termination:*** Backup Error *** -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] concurrent backups to one volume?
On Thu, Jun 25, 2009 at 12:42:30PM +0200, Andreas Schuldei wrote: hi! reading the documentation i understand that you should have several volumes for concurrent backups, on different devices/directories. (i work on disk for now.) However some people here on the list seem to be doing well with concurrent backups to only one volume. is that actually true or am i misunderstanding something? if so, how do i do that? i would like a reciept for that. what bacula version is required for that? Hallo Andreas, we're running two concurrent backups to one disk based volume using bacula 2.2.8. The volume grows to a maximum size of 400g (so we can fit two of them onto an lto4 for off-site storage), then the oldest one is recycled and used. So far, we haven't had any troubles restoring data with this setup. All I did in order to enable 2 concurrent jobs was to add the line Maximum concurrent jobs = 2 to the director definition. HTH, Uwe -- uwe.schuerk...@nionex.net phone: [+49] 5242.91 - 4740, fax:-69 72 Hauptsitz: Avenwedder Str. 55, D-33311 Guetersloh, Germany Registergericht Guetersloh HRB 4196, Geschaeftsfuehrer: Horst Gosewehr NIONEX ist ein Unternehmen der DirectGroup Germany www.directgroupgermany.de -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] concurrent backups to one volume?
Silver Salonen wrote: As one device supports only one job, you have to create separate devices for each job you want to be able to run concurrently. That isn't how I understand it. I am working on having multiple clients feeding files into a single tape drive at the same time and expect that chuncks of each job will be interleaved along the tape. It should be the same for disk files, if that ishow you configure it (say one file for each nights jobs). -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Fatal error: Job canceled because max start delay time exceeded
Hi every: Today in the morning I check my email and Bacula sending to me 139 emails with this error: Fatal error: Job canceled because max start delay time exceeded. What this means? How I can fix this? Regards, -- Ing. Reynier Pérez Mira -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] concurrent backups to one volume?
Andreas Schuldei wrote: hi! reading the documentation i understand that you should have several volumes for concurrent backups, on different devices/directories. (i work on disk for now.) However some people here on the list seem to be doing well with concurrent backups to only one volume. is that actually true or am i misunderstanding something? This is true, we use such a setup with tape based volumes. The point is that you should use spooling to disk in such a case to avoid interleaving of the jobs on tape. This is reported to prolong restore times. Since our clients cannot saturate our LTO drive we would need spooling anyhow and never tried without. Not sure if spooling would also be needed for disk based volumes though. if so, how do i do that? i would like a reciept for that. Please search the list, this was discussed several times. It involves adding MaximumConcurrentJobs directives in several places. what bacula version is required for that? Any not too ancient bacula version should do. We are using 2.2.8 Attila /andreas -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Fwd: concurrent backups to one volume?
On Thu, Jun 25, 2009 at 10:23 AM, John Drescherdresche...@gmail.com wrote: On Thu, Jun 25, 2009 at 9:34 AM, terrycter...@woa.com.au wrote: Silver Salonen wrote: As one device supports only one job, you have to create separate devices for each job you want to be able to run concurrently. That isn't how I understand it. I am working on having multiple clients feeding files into a single tape drive at the same time and expect that chuncks of each job will be interleaved along the tape. It should be the same for disk files, if that ishow you configure it (say one file for each nights jobs). As long as the pool is the same more than 1 job can concurrently write to the same volume. I have been doing that for years with tape and disk. However if you do want more than 1 pool then with disks its best to have multiple storage devices. I should have said more than 1 pool to operate concurrently with disk. John M. Drescher -- John M. Drescher -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] concurrent backups to one volume?
On Thu, Jun 25, 2009 at 9:34 AM, terrycter...@woa.com.au wrote: Silver Salonen wrote: As one device supports only one job, you have to create separate devices for each job you want to be able to run concurrently. That isn't how I understand it. I am working on having multiple clients feeding files into a single tape drive at the same time and expect that chuncks of each job will be interleaved along the tape. It should be the same for disk files, if that ishow you configure it (say one file for each nights jobs). As long as the pool is the same more than 1 job can concurrently write to the same volume. I have been doing that for years with tape and disk. However if you do want more than 1 pool then with disks its best to have multiple storage devices. John -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Antwort: Fatal error: Job canceled because max start delay time exceeded
Reynier Pérez Mira rper...@uci.cu wrote on 25.06.2009 15:51:46: Reynier Pérez Mira rper...@uci.cu 25.06.2009 15:51 Bitte antworten an rper...@uci.cu An bacula-users@lists.sourceforge.net Kopie Thema [Bacula-users] Fatal error: Job canceled because max start delay time exceeded Hi every: Today in the morning I check my email and Bacula sending to me 139 emails with this error: Fatal error: Job canceled because max start delay time exceeded. What this means? How I can fix this? Regards, -- Ing. Reynier Pérez Mira See Max Start Delay in the Job Resource. I guess one of your other Jobs took too long or hang. From the documentation: Max Start Delay = time The time specifies the maximum delay between the scheduled time and the actual start time for the Job. For example, a job can be scheduled to run at 1:00am, but because other jobs are running, it may wait to run. If the delay is set to 3600 (one hour) and the job has not begun to run by 2:00am, the job will be canceled. This can be useful, for example, to prevent jobs from running during day time hours. The default is 0 which indicates no limit. -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Antwort: Fatal error: Job canceled because max start delay time exceeded
c.kesch...@internet-mit-iq.de wrote: See Max Start Delay in the Job Resource. I guess one of your other Jobs took too long or hang. From the documentation: *Max Start Delay = time* The time specifies the maximum delay between the scheduled time and the actual start time for the Job. For example, a job can be scheduled to run at 1:00am, but because other jobs are running, it may wait to run. If the delay is set to 3600 (one hour) and the job has not begun to run by 2:00am, the job will be canceled. This can be useful, for example, to prevent jobs from running during day time hours. The default is 0 which indicates no limit. Thanks for your reply but this confuse me a bit. I have, for now, 40 clients and 40 Jobs one for each client. Every job start at 2:00 AM. My Director config is as follow: Director { Name = serverbacula-dir Description = Bacula Director Centro de Datos UCI DIRport = 9101 DirAddress = 10.128.50.11 QueryFile = /etc/bacula/query.sql WorkingDirectory = /var/bacula/working PidDirectory = /var/run Maximum Concurrent Jobs = 20 Password = some_password Messages = Daemon FD Connect Timeout = 5 min SD Connect Timeout = 5 min } So if I understood the Director resource and how Bacula works this configuration allow 20 Jobs running at the same time. Then a good configuration for Jobs could be set Max Start Delay to 2 hours? Give me a hint here because I'm lost. Regards, -- Ing. Reynier Pérez Mira -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Antwort: Re: Antwort: Fatal error: Job canceled because max start delay time exceeded
Reynier Pérez Mira rper...@uci.cu wrote on 25.06.2009 16:41:52: Reynier Pérez Mira rper...@uci.cu 25.06.2009 16:41 Bitte antworten an rper...@uci.cu An c.kesch...@internet-mit-iq.de Kopie bacula-users@lists.sourceforge.net Thema Re: Antwort: [Bacula-users] Fatal error: Job canceled because max start delay time exceeded c.kesch...@internet-mit-iq.de wrote: See Max Start Delay in the Job Resource. I guess one of your other Jobs took too long or hang. From the documentation: *Max Start Delay = time* The time specifies the maximum delay between the scheduled time and the actual start time for the Job. For example, a job can be scheduled to run at 1:00am, but because other jobs are running, it may wait to run. If the delay is set to 3600 (one hour) and the job has not begun to run by 2:00am, the job will be canceled. This can be useful, for example, to prevent jobs from running during day time hours. The default is 0 which indicates no limit. Thanks for your reply but this confuse me a bit. I have, for now, 40 clients and 40 Jobs one for each client. Every job start at 2:00 AM. My Director config is as follow: Director { Name = serverbacula-dir Description = Bacula Director Centro de Datos UCI DIRport = 9101 DirAddress = 10.128.50.11 QueryFile = /etc/bacula/query.sql WorkingDirectory = /var/bacula/working PidDirectory = /var/run Maximum Concurrent Jobs = 20 Password = some_password Messages = Daemon FD Connect Timeout = 5 min SD Connect Timeout = 5 min } So if I understood the Director resource and how Bacula works this configuration allow 20 Jobs running at the same time. Then a good configuration for Jobs could be set Max Start Delay to 2 hours? Give me a hint here because I'm lost. Regards, -- Ing. Reynier Pérez Mira Oh you have 40 clients/jobs. I misread before. Well there are surely some Jobs that don't finish in the two hours. Even if 15 Jobs finish and 5 don't there are then still 5 jobs waiting to execute.-- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Antwort: Re: Antwort: Fatal error: Job canceled because max start delay time exceeded
Reynier Pérez Mira rper...@uci.cu wrote on 25.06.2009 16:41:52: Reynier Pérez Mira rper...@uci.cu 25.06.2009 16:41 Bitte antworten an rper...@uci.cu An c.kesch...@internet-mit-iq.de Kopie bacula-users@lists.sourceforge.net Thema Re: Antwort: [Bacula-users] Fatal error: Job canceled because max start delay time exceeded c.kesch...@internet-mit-iq.de wrote: See Max Start Delay in the Job Resource. I guess one of your other Jobs took too long or hang. From the documentation: *Max Start Delay = time* The time specifies the maximum delay between the scheduled time and the actual start time for the Job. For example, a job can be scheduled to run at 1:00am, but because other jobs are running, it may wait to run. If the delay is set to 3600 (one hour) and the job has not begun to run by 2:00am, the job will be canceled. This can be useful, for example, to prevent jobs from running during day time hours. The default is 0 which indicates no limit. Thanks for your reply but this confuse me a bit. I have, for now, 40 clients and 40 Jobs one for each client. Every job start at 2:00 AM. My Director config is as follow: Director { Name = serverbacula-dir Description = Bacula Director Centro de Datos UCI DIRport = 9101 DirAddress = 10.128.50.11 QueryFile = /etc/bacula/query.sql WorkingDirectory = /var/bacula/working PidDirectory = /var/run Maximum Concurrent Jobs = 20 Password = some_password Messages = Daemon FD Connect Timeout = 5 min SD Connect Timeout = 5 min } So if I understood the Director resource and how Bacula works this configuration allow 20 Jobs running at the same time. Then a good configuration for Jobs could be set Max Start Delay to 2 hours? Give me a hint here because I'm lost. Regards, -- Ing. Reynier Pérez Mira Do any jobs have a higher/lower priority? If so, the other jobs will wait until these finish (you can somehow configure bacula to run jobs with different priorities at the same time iirc, but I would have to look that up in the docs)-- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Fwd: concurrent backups to one volume?
On Thursday 25 June 2009 17:24:37 John Drescher wrote: On Thu, Jun 25, 2009 at 10:23 AM, John Drescherdresche...@gmail.com wrote: On Thu, Jun 25, 2009 at 9:34 AM, terrycter...@woa.com.au wrote: Silver Salonen wrote: As one device supports only one job, you have to create separate devices for each job you want to be able to run concurrently. That isn't how I understand it. I am working on having multiple clients feeding files into a single tape drive at the same time and expect that chuncks of each job will be interleaved along the tape. It should be the same for disk files, if that ishow you configure it (say one file for each nights jobs). As long as the pool is the same more than 1 job can concurrently write to the same volume. I have been doing that for years with tape and disk. However if you do want more than 1 pool then with disks its best to have multiple storage devices. I should have said more than 1 pool to operate concurrently with disk. John M. Drescher Yes, in this case we are about to ask ourselves what are pools - to my mind pools are collections of backup-files and policies about how to overwrite these files. Eg. if we want to do ordinary Grandfather-Father-Son rotation, we create 3 different pools for every job - for full, differential and incremental backups. And as we define them, we have to define separate devices for them too. Using only one pool with the whole backup-disk doesn't make sense to me, because managing these backups would be extremely limited, wouldn't it? PS. The limit to be able to write only one job to one disk-based device has been a bizzare limit that just complicates the configuration, I still don't understand why we have this limit in disk-based backups (the claim Bacula uses disks as tapes is just as bizzare). -- Silver -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Fwd: concurrent backups to one volume?
On Thursday 25 June 2009 18:10:35 John Drescher wrote: PS. The limit to be able to write only one job to one disk-based device has been a bizzare limit that just complicates the configuration, I still don't understand why we have this limit in disk-based backups (the claim Bacula uses disks as tapes is just as bizzare). There is no such limit. If you want more than one pool to write concurrently have more than 1 storage device. With disks you can have as many as you want. They can all point to the same physical storage location. I meant the configuration limit - that I can't configure one device to accept multiple jobs concurrently. -- Silver -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Fwd: concurrent backups to one volume?
Silver Salonen schrieb: PS. The limit to be able to write only one job to one disk-based device has been a bizzare limit that just complicates the configuration, I still don't understand why we have this limit in disk-based backups (the claim Bacula uses disks as tapes is just as bizzare). I seem to be writing to disk based volumes just fine with multiple (5+) concurrent jobs. Maybe i am misunderstanding something, but the problem isnt N jobs to 1 (disk based) volume, it is 1-N jobs to M (1) volumes at the same time which doesnt work. Also, one pool can only use one volume at one time for one job, and one job can only use one SD at one time, so that might limit you too. Concurrent jobs (except for me currently, and only to tape) work quite nice. I personally opted to use spooling, in order to not have jobs interleaved on tape. But speaking of configuration complication, for me it complicates things that valid jobs (for comparing with Full, Diffs and Incrementals) are generated out of {Client,Fileset}and there is no way to tell it to use {Client,Fileset,Storage} or {Client,Fileset,Pool}.. for creating offsite backups i basically keep 2 identical filesets for every class of client because i want / need two unrelated backups on 2 different media types (no, copy jobs wont do because they are 2 different SDs).. Anyways, i am rambling, but multiple concurrent jobs to one disk volume work fine, as long as your jobs use the same pool and are allowed to use the same volume. If you try to use different pools ( - different volumes) for different jobs, then yes, they will wait till the SD can mount a volume in the second pool. -- Christian Gaul otop AG D-55116 Mainz Rheinstraße 105-107 Fon: 06131.5763.310 Fax: 06131.5763.500 E-Mail: christian.g...@otop.de Internet: www.otop.de Vorsitzender des Aufsichtsrats: Christof Glasmacher Vorstand: Dirk Flug Registergericht: Amtsgericht Mainz Handelsregister: HRB 7647 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] estimate and actual job differ
On Friday 05 June 2009 16:36:16 Silver Salonen wrote: On Thursday 04 June 2009 10:53:03 Christian Gaul wrote: Silver Salonen schrieb: On Thursday 04 June 2009 10:34:36 Christian Gaul wrote: Silver Salonen schrieb: Hi. I'm trying to run incremental job of a restored fileset (having mtimeonly=yes). When I check its estimate, it shows correctly only new files that have been created/modified since restoration. But when I run the actual job, all the files are included in backup. The server is 3.0.0 on FreeBSD, client is 3.0.1 on Windows XP. May it be because all the folders' (but not files') mtime is the date of restoration? And I wonder what is the latter caused by? Since you modified the fileset to add mtimeonly=yes, did you also add Ignore FileSet Changes=yes? If not, your next backup will default to a Full because the fileset doesnt match to the one your last Full was made with. Yes, I also have Ignore FileSet Changes=yes, sorry I didn't mention it. And if I didn't have it, estimate would show full too, wouldn't it? I dont know if estimate honors that or just takes what you give it. I personally only use estimate when making new filesets (to see if my excludes work correctly). And since using estimate with LVM Snapshots doesnt work anyways because they are not mounted for an estimate job, most of my filesets would show 0 files anyways. Sorry i cant help with that. I suppose estimate does honor the option (Ignore FileSet Changes=yes), because the problem is not in estimate job's files, but rather in actual job's files - the actual job just wants to back up all the files, but it should not. It seems that actual job (or the client of the job) doesn't honor the mtimeonly=yes - it worked OK with my FreeBSD-client when I tested the similar situation. -- Silver I submitted a bug report for that: http://bugs.bacula.org:80/view.php?id=1318 -- Silver -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Fwd: concurrent backups to one volume?
On Thursday 25 June 2009 18:28:28 Christian Gaul wrote: Silver Salonen schrieb: PS. The limit to be able to write only one job to one disk-based device has been a bizzare limit that just complicates the configuration, I still don't understand why we have this limit in disk-based backups (the claim Bacula uses disks as tapes is just as bizzare). I seem to be writing to disk based volumes just fine with multiple (5+) concurrent jobs. Maybe i am misunderstanding something, but the problem isnt N jobs to 1 (disk based) volume, it is 1-N jobs to M (1) volumes at the same time which doesnt work. Also, one pool can only use one volume at one time for one job, and one job can only use one SD at one time, so that might limit you too. Concurrent jobs (except for me currently, and only to tape) work quite nice. I personally opted to use spooling, in order to not have jobs interleaved on tape. But speaking of configuration complication, for me it complicates things that valid jobs (for comparing with Full, Diffs and Incrementals) are generated out of {Client,Fileset}and there is no way to tell it to use {Client,Fileset,Storage} or {Client,Fileset,Pool}.. for creating offsite backups i basically keep 2 identical filesets for every class of client because i want / need two unrelated backups on 2 different media types (no, copy jobs wont do because they are 2 different SDs).. Anyways, i am rambling, but multiple concurrent jobs to one disk volume work fine, as long as your jobs use the same pool and are allowed to use the same volume. If you try to use different pools ( - different volumes) for different jobs, then yes, they will wait till the SD can mount a volume in the second pool. I meant the limitation from the configuration point of view - you cannot configure a device to accept multiple jobs concurrently. If you want to be able to actually do it, you have to hack the configuration - to show one actual device as different devices. -- Silver -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Fwd: concurrent backups to one volume?
Silver Salonen schrieb: On Thursday 25 June 2009 18:28:28 Christian Gaul wrote: Silver Salonen schrieb: PS. The limit to be able to write only one job to one disk-based device has been a bizzare limit that just complicates the configuration, I still don't understand why we have this limit in disk-based backups (the claim Bacula uses disks as tapes is just as bizzare). I seem to be writing to disk based volumes just fine with multiple (5+) concurrent jobs. Maybe i am misunderstanding something, but the problem isnt N jobs to 1 (disk based) volume, it is 1-N jobs to M (1) volumes at the same time which doesnt work. Also, one pool can only use one volume at one time for one job, and one job can only use one SD at one time, so that might limit you too. Concurrent jobs (except for me currently, and only to tape) work quite nice. I personally opted to use spooling, in order to not have jobs interleaved on tape. But speaking of configuration complication, for me it complicates things that valid jobs (for comparing with Full, Diffs and Incrementals) are generated out of {Client,Fileset}and there is no way to tell it to use {Client,Fileset,Storage} or {Client,Fileset,Pool}.. for creating offsite backups i basically keep 2 identical filesets for every class of client because i want / need two unrelated backups on 2 different media types (no, copy jobs wont do because they are 2 different SDs).. Anyways, i am rambling, but multiple concurrent jobs to one disk volume work fine, as long as your jobs use the same pool and are allowed to use the same volume. If you try to use different pools ( - different volumes) for different jobs, then yes, they will wait till the SD can mount a volume in the second pool. I meant the limitation from the configuration point of view - you cannot configure a device to accept multiple jobs concurrently. If you want to be able to actually do it, you have to hack the configuration - to show one actual device as different devices. I think i understand what you mean, but you actually can accept multiple jobs to the same device.. just not to different pools. But you are right, since it's disk volumes, one file based SD could act like a virtual tape changer with unlimited slots and drives... i guess you probably could create a virtual changer with like 20 drives, but you are right, that is kind of hack-isch. If i find time i might experiment with the current possibilities of virtual libraries and see if i can work something out that doesnt look like a hack. -- Christian Gaul otop AG D-55116 Mainz Rheinstraße 105-107 Fon: 06131.5763.310 Fax: 06131.5763.500 E-Mail: christian.g...@otop.de Internet: www.otop.de Vorsitzender des Aufsichtsrats: Christof Glasmacher Vorstand: Dirk Flug Registergericht: Amtsgericht Mainz Handelsregister: HRB 7647 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Fwd: concurrent backups to one volume?
There is no such limit. If you want more than one pool to write concurrently have more than 1 storage device. With disks you can have as many as you want. They can all point to the same physical storage location. I meant the configuration limit - that I can't configure one device to accept multiple jobs concurrently. I do this every single day at home. 5 jobs concurrently write to the same exact volume. -- John M. Drescher -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Problems restoring an Exchange plugin backup
Update: I think I found a restore setup that actually restores the backup. The backup gets cut short when the DB is activated in Exchange and this also crashes the Bacula FD on the target machine. The end is Error: HrESERestoreComplete failed with error 0xc7ff1004 - Unknown error. at which point the FD crashes. When I try to mount the store (which is in fact 8GB in size, the same as it was before the backup - making me think the data should be there), I get an error and the event logs in Windows show this: Information Store (5212) Recovery Storage Group: Attempted to attach database 'C:\Program Files\Exchsrvr\Recovery Storage Group\Mailbox Store (AXMAIL).edb' but it is a database restored from a backup set on which hard recovery was not started or did not complete successfully. So what is still going wrong here? The complete transcript is down below (bconsole): Connecting to Director axnet:9101 1000 OK: axnet-dir Version: 3.0.1 (30 April 2009) Enter a period to cancel a command. *restore Automatically selected Catalog: MyCatalog Using Catalog MyCatalog First you select one or more JobIds that contain files to be restored. You will be presented several methods of specifying the JobIds. Then you will be allowed to select which files from those JobIds are to be restored. To select the JobIds, you have the following choices: 1: List last 20 Jobs run 2: List Jobs where a given File is saved 3: Enter list of comma separated JobIds to select 4: Enter SQL list command 5: Select the most recent backup for a client 6: Select backup for a client before a specified time 7: Enter a list of files to restore 8: Enter a list of files to restore before a specified time 9: Find the JobIds of the most recent backup for a client 10: Find the JobIds for a backup for a client before a specified time 11: Enter a list of directories to restore for found JobIds 12: Cancel Select item: (1-12): 5 Defined Clients: ... 4: axmail-fd ... 10: axemail-fd Select the Client (1-10): 4 The defined FileSet resources are: 1: AXMAIL Full Data Set 2: Exchange Select FileSet resource (1-2): 2 +---+---+--+---+-+---+ | JobId | Level | JobFiles | JobBytes | StartTime | VolumeName| +---+---+--+---+-+---+ |90 | F | 13 | 5,313,968,371 | 2009-06-24 15:36:10 | Deventer_Exchange_Backup_0013 | |90 | F | 13 | 5,313,968,371 | 2009-06-24 15:36:10 | Deventer_Exchange_Backup_0014 | |91 | I |5 | 2,671,174 | 2009-06-24 17:28:25 | Deventer_Exchange_Backup_0014 | |92 | I |5 | 233,882 | 2009-06-24 18:00:01 | Deventer_Exchange_Backup_0014 | | 118 | I | 17 |40,099,025 | 2009-06-25 18:00:02 | Deventer_Exchange_Backup_0014 | +---+---+--+---+-+---+ You have selected the following JobIds: 90,91,92,118 Building directory tree for JobId(s) 90,91,92,118 ... 24 files inserted into the tree. You are now entering file selection mode where you add (mark) and remove (unmark) files to be restored. No files are initially added, unless you used the all keyword on the command line. Enter done to leave this mode. cwd is: / $ mark * 29 files marked. $ cd @EXCHANGE/Microsoft Information Store/First Storage Group cwd is: /@EXCHANGE/Microsoft Information Store/First Storage Group/ $ unmark Public* 4 files unmarked. $ lsmark *C:\Program Files\Exchsrvr\mdbdata\E0002FC5.log *C:\Program Files\Exchsrvr\mdbdata\E0002FC6.log *C:\Program Files\Exchsrvr\mdbdata\E0002FC7.log *C:\Program Files\Exchsrvr\mdbdata\E0002FC8.log *C:\Program Files\Exchsrvr\mdbdata\E0002FC9.log *C:\Program Files\Exchsrvr\mdbdata\E0002FCA.log *C:\Program Files\Exchsrvr\mdbdata\E0002FCB.log *C:\Program Files\Exchsrvr\mdbdata\E0002FCC.log *C:\Program Files\Exchsrvr\mdbdata\E0002FCD.log *C:\Program Files\Exchsrvr\mdbdata\E0002FCE.log *C:\Program Files\Exchsrvr\mdbdata\E0002FCF.log *C:\Program Files\Exchsrvr\mdbdata\E0002FD0.log *C:\Program Files\Exchsrvr\mdbdata\E0002FD1.log *C:\Program Files\Exchsrvr\mdbdata\E0002FD2.log *C:\Program Files\Exchsrvr\mdbdata\E0002FD3.log *C:\Program Files\Exchsrvr\mdbdata\E0002FD4.log *C:\Program Files\Exchsrvr\mdbdata\E0002FD5.log *C:\Program Files\Exchsrvr\mdbdata\E0002FD6.log *Mailbox Store (AXMAIL)/ *C:\Program Files\Exchsrvr\mdbdata\priv1.edb *C:\Program Files\Exchsrvr\mdbdata\priv1.stm *DatabaseBackupInfo $ done Bootstrap records written to /var/bacula/axnet-dir.restore.17.bsr The job will require the following Volume(s) Storage(s)SD Device(s) === Deventer_Exchange_Backup_ File FileStorage 25
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
[Bacula-users] Overriding Job and/or File Retention Periods?
Thanks for the suggestions Dirk. I was afraid that it would come down to scripting; being a temp Student admin I'm trying to keep things as straightforward as possible for the next person who has to deal with it. I was hoping I had missed something and that the Job resource could also set the job retention period. Thank You :) Tom +-- |This was sent by tomisom.s...@gmail.com via Backup Central. |Forward SPAM to ab...@backupcentral.com. +-- -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] concurrent backups to one volume?
On 25.06.2009 14:10, Silver Salonen wrote: On Thursday 25 June 2009 13:42:30 Andreas Schuldei wrote: hi! reading the documentation i understand that you should have several volumes for concurrent backups, on different devices/directories. (i work on disk for now.) However some people here on the list seem to be doing well with concurrent backups to only one volume. is that actually true or am i misunderstanding something? if so, how do i do that? i would like a reciept for that. what bacula version is required for that? /andreas Hello. As one device supports only one job, you have to create separate devices for each job you want to be able to run concurrently. I´m not sure if you are talking about the same thing, but how should it possible to get interleaved volume blocks (where bacula must sort and take longer) if you never write to the same device (with or without spooling)? Quote from the current bacula-manual: The Volume format becomes more complicated with multiple simultaneous jobs, consequently, restores may take longer if Bacula must sort through interleaved volume blocks from multiple simultaneous jobs. This can be avoided by having each simultaneous job write to a different volume or by using data spooling, which will first spool the data to disk simultaneously, then write one spool file at a time to the volume thus avoiding excessive interleaving of the different job blocks. Previous to 3.x I run concurrent jobs and have just one device (a tapedrive in an autoloader) but a huge disk-space for spooling. And it worked well. However I (and I´m not the only one) got troubles with concurrent jobs on 3.x now (files mismatch). I´m sure I can avoid that by using different backup devices but I just have one tapedrive built in the loader so I would not do that. Greetings, user100 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Backup space quotas
Hello, I'm trying to determine if there's a way to define space quotas on a per host basis in Bacula. I figured there might be a way to do it by limiting the size of volumes, number of volumes within a pool, and assigning each host it's own pool. I read in an old thread that this goes against the design principles of Bacula, but I could not find anything definitive. Any direct assistance or ideas to explore would be greatly appreciated. Thank you! -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Backup space quotas
I'm trying to determine if there's a way to define space quotas on a per host basis in Bacula. I figured there might be a way to do it by limiting the size of volumes, number of volumes within a pool, and assigning each host it's own pool. That sounds fine. I would do exactly that if I had this requirement. John -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Problems restoring an Exchange plugin backup
Another update: I found out that the database is in fact in a Dirty Shutdown state (eseutil.exe told me that) - hence it won't work (mount). I found a discussion about this here: http://www.mail-archive.com/bacula-users@lists.sourceforge.net/msg30912.html But it seems like there was no solution for the problem. Am I right to conclude that the database is in fact copied while active and as such in the dirt state? Shouldn't the plugin handle this? Regards, Berend Dekens Op 06/25/09 19:37, Berend Dekens schreef: Update: I think I found a restore setup that actually restores the backup. The backup gets cut short when the DB is activated in Exchange and this also crashes the Bacula FD on the target machine. The end is Error: HrESERestoreComplete failed with error 0xc7ff1004 - Unknown error. at which point the FD crashes. When I try to mount the store (which is in fact 8GB in size, the same as it was before the backup - making me think the data should be there), I get an error and the event logs in Windows show this: Information Store (5212) Recovery Storage Group: Attempted to attach database 'C:\Program Files\Exchsrvr\Recovery Storage Group\Mailbox Store (AXMAIL).edb' but it is a database restored from a backup set on which hard recovery was not started or did not complete successfully. So what is still going wrong here? snip -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Fwd: concurrent backups to one volume?
Silver Salonen wrote: Eg. if we want to do ordinary Grandfather-Father-Son rotation, we create 3 different pools for every job - for full, differential and incremental backups. That is not how I understand GFS system, although it is a possibility. I understand it as Full, plus (incremental OR differential). So important clients (like secretary's machine) receive a full backup each week and a differential (all changed files since full backup) nightly so that in the need for recovery, it would just be a process involving two tape/disk(?) for a full recovery. OTOH, I might do a differential (all changed files since last backup, full or diff or inc) on something on something with humungous amount of file changes and non-core/non-critical files to simply keep the backup window small. The trade off is that every tape/disk since the full backup would need to be processed for a full client recovery. GFS comes from having multiple complete BACKUPS, i.e. dated versions. This makes it a real backup system. Using only one pool with the whole backup-disk doesn't make sense to me, because managing these backups would be extremely limited, wouldn't it? If you want to run your backups like that, then use a raid array. You only have a point in the rare occassions where you have small backups and monstrous drives. The critical point about a real backup system is that it is not just a file copy, but a secure,protected file copy that can not be degraded. Writing a whole serious of jobss to one drive that sits in the system full time is not a proper backup system. -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Migration Job pruging destination volumes!!
I've set-up a migration job to migrate jobs from one set of tape volumes to disk volumes. I've configured the destination pool to use the volume once and have a retention period of 2 months. For some reason when the migration job completes and gets to the next queued migration job it marks the destination volume as purged and then overwrites the contents with the next job. The Jobs from the source pool are being marked as purged, so the data is going straight into the bit bucket! I'm using version 2.4.4 from Debian and here is the pool and job portion of my conf files. Pool { Name = 454FLX Pool Type = Backup AutoPrune = yes Storage = Neo8000-LTO4 VolumeRetention = 3 years Recycle = yes Next Pool = DD-454FLX } Pool { Name = DD-454FLX Pool Type = Backup LabelFormat = 454FLX- Recycle = yes AutoPrune = yes Storage = DD-454FLX Volume Retention = 2 months Use Volume Once = yes } Job { Name = Migrate_454FLX Type = Migrate Level = Full Client = 454datarig-fd FileSet = FULL Windows Messages = Standard Pool = 454FLX Maximum Concurrent Jobs = 4 Selection Type = Volume Selection Pattern = .*L4 } Here is a piece of the output that is confirming my bit bucket suspicion: 25-Jun 17:50 babacula-dir JobId 37438: Start Migration JobId 37438, Job=Migrate_454FLX.2009-06-25_15.15.54.27 25-Jun 17:50 babacula-dir JobId 37438: There are no more Jobs associated with Volume 454FLX-0169. Marking it purged. 25-Jun 17:50 babacula-dir JobId 37438: All records pruned from Volume 454FLX-0169; marking it Purged 25-Jun 17:50 babacula-dir JobId 37438: Recycled volume 454FLX-0169 25-Jun 17:50 babacula-dir JobId 37438: Using Device DD-454FLX 25-Jun 17:50 lsbacsd0-sd JobId 37438: Ready to read from volume 02L4 on device Drive-2 (/dev/tape/drive2). 25-Jun 17:50 lsbacsd0-sd JobId 37438: Recycled volume 454FLX-0169 on device DD-454FLX (/backup/pools/454FLX), all previous data lost. 25-Jun 17:51 babacula-dir JobId 37438: Volume used once. Marking Volume 454FLX-0169 as Used. 25-Jun 17:50 lsbacsd0-sd JobId 37438: Forward spacing Volume 02L4 to file:block 418:0. So, two questions. 1. What am I doing wrong? 2. Is there an easy way to unpurge the jobs on the tape since they have not been recycled, or do I have to run bscan on them? Thanks, Robert LeBlanc Life Sciences Undergraduate Education Computer Support Brigham Young University -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Problems restoring an Exchange plugin backup
Another update: I found out that the database is in fact in a Dirty Shutdown state (eseutil.exe told me that) - hence it won't work (mount). I found a discussion about this here: http://www.mail-archive.com/bacula-users@lists.sourceforge.net/msg30912. html But it seems like there was no solution for the problem. Am I right to conclude that the database is in fact copied while active and as such in the dirt state? Shouldn't the plugin handle this? How urgent is it to get this fixed? As far as I can tell you are doing everything right... I'll look up those error messages. James -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] concurrent backups to one volume?
Attila Fülöp wrote: Andreas Schuldei wrote: hi! reading the documentation i understand that you should have several volumes for concurrent backups, on different devices/directories. (i work on disk for now.) However some people here on the list seem to be doing well with concurrent backups to only one volume. is that actually true or am i misunderstanding something? Yes. The hard part is getting it configured in ALL the needed places. anything up to four places in a couple of file (dir Sd) and in the client) as well. The line is maximum concurrent jobs = N where N= number you want. Note, there are some default settings in SD FD that allow for ten(?) and only require a DIR change to work. Caution; do not run any other job at the same time as catalog dumps (full catalog) This is true, we use such a setup with tape based volumes. The point is that you should use spooling to disk in such a case to avoid interleaving of the jobs on tape. AFAIUI, if the jobs is smaller than the spool size, then the jobs will not be interleaved. If the job is larger than the spool size, then segments will be interleaved with other running jobs. -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Fwd: concurrent backups to one volume?
On Thursday 25 June 2009 19:41:10 John Drescher wrote: There is no such limit. If you want more than one pool to write concurrently have more than 1 storage device. With disks you can have as many as you want. They can all point to the same physical storage location. I meant the configuration limit - that I can't configure one device to accept multiple jobs concurrently. I do this every single day at home. 5 jobs concurrently write to the same exact volume. My original claim was made in the context of disk-based backups (ie. multiple pools as I explained in the same message). Using the same exact volume (or pool) with disk-based backup-system is quite a big limitation to my mind. -- Silver -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Fwd: concurrent backups to one volume?
On Friday 26 June 2009 02:07:58 terryc wrote: Silver Salonen wrote: Eg. if we want to do ordinary Grandfather-Father-Son rotation, we create 3 different pools for every job - for full, differential and incremental backups. That is not how I understand GFS system, although it is a possibility. I understand it as Full, plus (incremental OR differential). So important clients (like secretary's machine) receive a full backup each week and a differential (all changed files since full backup) nightly so that in the need for recovery, it would just be a process involving two tape/disk(?) for a full recovery. OTOH, I might do a differential (all changed files since last backup, full or diff or inc) on something on something with humungous amount of file changes and non-core/non-critical files to simply keep the backup window small. The trade off is that every tape/disk since the full backup would need to be processed for a full client recovery. GFS comes from having multiple complete BACKUPS, i.e. dated versions. This makes it a real backup system. OK, yes.. you may do it as this too, but the point in this context was that we need multiple pools. In my case I need one pool for full backups, one for differentials and one for incrementals. In your case you need 2 pools: one for fulls and one for differentials. Using only one pool with the whole backup-disk doesn't make sense to me, because managing these backups would be extremely limited, wouldn't it? If you want to run your backups like that, then use a raid array. You only have a point in the rare occassions where you have small backups and monstrous drives. The critical point about a real backup system is that it is not just a file copy, but a secure,protected file copy that can not be degraded. Writing a whole serious of jobss to one drive that sits in the system full time is not a proper backup system. Well.. I AM using RAID array everywhere in backup-systems and as it's so much more cost effective than using tapes, we just hope we can detect any soon-to- failure storage soon enough :) -- Silver -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users