Re: [Bacula-users] continuing a failed job
On Wed, Jan 06, 2010 at 12:52:44PM -0500, Phil Stracchino wrote: My first inclination is to say that if the network connection to the machine is sufficiently unstable that you can't complete a full backup, you probably shouldn't be trying to back it up over the network. However, there are ways to work around the problem. You could also try rsync'ing (rsync -av u...@remotehost:/ /backup/remote_host/) to a local filesystem and then backing that copy up locally. Should rsync abort it will skip the files which are already present up to date on the local copy. HTH, Uwe -- uwe.schuerk...@nionex.net fon: [+49] 5242.91 - 4740, fax:-69 72 Hauptsitz: Avenwedder Str. 55, D-33311 Gütersloh, Germany Registergericht Gütersloh HRB 4196, Geschäftsführer: H. Gosewehr, D. Suda NIONEX ist ein Unternehmen der DirectGroup Germany www.directgroupgermany.de -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] continuing a failed job
On Thursday 07 January 2010 10:45:27 Uwe Schuerkamp wrote: On Wed, Jan 06, 2010 at 12:52:44PM -0500, Phil Stracchino wrote: My first inclination is to say that if the network connection to the machine is sufficiently unstable that you can't complete a full backup, you probably shouldn't be trying to back it up over the network. However, there are ways to work around the problem. You could also try rsync'ing (rsync -av u...@remotehost:/ /backup/remote_host/) to a local filesystem and then backing that copy up locally. Should rsync abort it will skip the files which are already present up to date on the local copy. It's a Windows workstation and I've set up the system where the person can just click a shortcut that activates his job and he gets the notification about the job by e-mail later. So rsync is not an option. -- Silver -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] continuing a failed job
Uwe Schuerkamp wrote: On Wed, Jan 06, 2010 at 12:52:44PM -0500, Phil Stracchino wrote: My first inclination is to say that if the network connection to the machine is sufficiently unstable that you can't complete a full backup, you probably shouldn't be trying to back it up over the network. However, there are ways to work around the problem. You could also try rsync'ing (rsync -av u...@remotehost:/ /backup/remote_host/) to a local filesystem and then backing that copy up locally. Should rsync abort it will skip the files which are already present up to date on the local copy. You know, I believe that's exactly what I suggested in the rest of the message that you just quoted the beginning of above. :) -- 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, Free Stater It's not the years, it's the mileage. -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] continuing a failed job
It's a Windows workstation and I've set up the system where the person can just click a shortcut that activates his job and he gets the notification about the job by e-mail later. So rsync is not an option. Why not? I do a lot of bacula based backups from windows machines that utilize cwrsync as a starting point for staging for this very reason. -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] continuing a failed job
On Thursday 07 January 2010 15:41:15 Phil Stracchino wrote: Silver Salonen wrote: On Thursday 07 January 2010 10:45:27 Uwe Schuerkamp wrote: On Wed, Jan 06, 2010 at 12:52:44PM -0500, Phil Stracchino wrote: My first inclination is to say that if the network connection to the machine is sufficiently unstable that you can't complete a full backup, you probably shouldn't be trying to back it up over the network. However, there are ways to work around the problem. You could also try rsync'ing (rsync -av u...@remotehost:/ /backup/remote_host/) to a local filesystem and then backing that copy up locally. Should rsync abort it will skip the files which are already present up to date on the local copy. It's a Windows workstation and I've set up the system where the person can just click a shortcut that activates his job and he gets the notification about the job by e-mail later. So rsync is not an option. Consider DeltaCopy (http://www.aboutmyip.com/AboutMyXApp/DeltaCopy.jsp), which is basically rsync for Windows in a Windows-friendly wrapper. It's free, GPL'd, and will connect to a Unix rsyncd. Thus, rsync *is* an option. Well, if I plan to use it along with Bacula, a client needs another amount of data-space in a backup-server (and a chroot'ed user for rsync) - user would then rsync his data into the server and Bacula would back it up every night, right? Sounds complicated for a wider use. I still think that if there was a way for marking a failed job as successful, it would be just about enough. -- Silver -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] continuing a failed job
Silver Salonen wrote: On Thursday 07 January 2010 10:45:27 Uwe Schuerkamp wrote: On Wed, Jan 06, 2010 at 12:52:44PM -0500, Phil Stracchino wrote: My first inclination is to say that if the network connection to the machine is sufficiently unstable that you can't complete a full backup, you probably shouldn't be trying to back it up over the network. However, there are ways to work around the problem. You could also try rsync'ing (rsync -av u...@remotehost:/ /backup/remote_host/) to a local filesystem and then backing that copy up locally. Should rsync abort it will skip the files which are already present up to date on the local copy. It's a Windows workstation and I've set up the system where the person can just click a shortcut that activates his job and he gets the notification about the job by e-mail later. So rsync is not an option. Consider DeltaCopy (http://www.aboutmyip.com/AboutMyXApp/DeltaCopy.jsp), which is basically rsync for Windows in a Windows-friendly wrapper. It's free, GPL'd, and will connect to a Unix rsyncd. Thus, rsync *is* an option. -- 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, Free Stater It's not the years, it's the mileage. -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] continuing a failed job
On Thursday 07 January 2010 14:25:04 Joseph L. Casale wrote: It's a Windows workstation and I've set up the system where the person can just click a shortcut that activates his job and he gets the notification about the job by e-mail later. So rsync is not an option. Why not? I do a lot of bacula based backups from windows machines that utilize cwrsync as a starting point for staging for this very reason. What do you think about what I wrote in my previous e-mail (about this approach not being a very simple one in terms of wider use)? -- Silver -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] continuing a failed job
Silver Salonen wrote: On Thursday 07 January 2010 15:41:15 Phil Stracchino wrote: Thus, rsync *is* an option. Well, if I plan to use it along with Bacula, a client needs another amount of data-space in a backup-server (and a chroot'ed user for rsync) - user would then rsync his data into the server and Bacula would back it up every night, right? Sounds complicated for a wider use. A little more complicated, but which is better: A slightly more complicated (to set up in the first place) solution that works, and - as far as I can tell from the page I linked - works without any user action required once set up; or a solution that's simpler to set up, but requires a manual user action, and demonstrably doesn't work? Any system that works beats one which demonstrably doesn't. Your client's network connection is demonstrably not stable enough to use Bacula directly. Bacula itself is not designed to work around a connection that unstable. The rsync protocol is, and Bacula will happily and reliably back up the rsync'd local mirror. I still think that if there was a way for marking a failed job as successful, it would be just about enough. I think I can safely say that this will never be a Bacula feature, and nor should it. If you create the capability to mark failed backups as successful, you can no longer trust your backups, because you can never know for sure that any set of backups which contains one or more jobs which failed, but were then marked as successful, actually contains ANY complete and consistent image of the filesystem. -- 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, Free Stater It's not the years, it's the mileage. -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] continuing a failed job
On Thursday 07 January 2010 16:06:50 Phil Stracchino wrote: Silver Salonen wrote: On Thursday 07 January 2010 15:41:15 Phil Stracchino wrote: Thus, rsync *is* an option. Well, if I plan to use it along with Bacula, a client needs another amount of data-space in a backup-server (and a chroot'ed user for rsync) - user would then rsync his data into the server and Bacula would back it up every night, right? Sounds complicated for a wider use. A little more complicated, but which is better: A slightly more complicated (to set up in the first place) solution that works, and - as far as I can tell from the page I linked - works without any user action required once set up; or a solution that's simpler to set up, but requires a manual user action, and demonstrably doesn't work? Well, my approach can easily be set up without requiring manual user interaction too. And what doesn't work, is only the first full backup - incrementals would be no problem once the full is done. Not the solution, right? I still think that if there was a way for marking a failed job as successful, it would be just about enough. I think I can safely say that this will never be a Bacula feature, and nor should it. If you create the capability to mark failed backups as successful, you can no longer trust your backups, because you can never know for sure that any set of backups which contains one or more jobs which failed, but were then marked as successful, actually contains ANY complete and consistent image of the filesystem. I'm not saying it should be a feature. I'm asking is there a way to do it, eg. by changing smth in database, rescanning the volume etc. What could also help and what looks more like a feature, is ability to run a job from a failed job (using a failed full as basis for incremental), which is exactly what I want. -- Silver -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] continuing a failed job
Am Wed, 06 Jan 2010 11:08:17 +0200 schrieb Silver Salonen: Hi. Has anyone figured out how to continue a failed job? I have a client that has gigabytes of data, but very fragile internet connection, so it's almost impossible to get a normal full backup job. I thought I'd do it with VirtualFull job to create a successful job out of failed one and then use it as basis for incremental, but unfortunately VirtualFull requires a previous successful job as well (duh). Is there a way to mark a job successful? you can't continue a failed job. this is a known problem with unstable internet connections. maybe you can work around with openvpn (or something like that) to simply hide short outages to bacula. - Thomas -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] continuing a failed job
Thomas Mueller tho...@chaschperli.ch kirjoitti viestissä news:hi2ffj$ga...@ger.gmane.org... Am Wed, 06 Jan 2010 11:08:17 +0200 schrieb Silver Salonen: Hi. Has anyone figured out how to continue a failed job? I have a client that has gigabytes of data, but very fragile internet connection, so it's almost impossible to get a normal full backup job. I thought I'd do it with VirtualFull job to create a successful job out of failed one and then use it as basis for incremental, but unfortunately VirtualFull requires a previous successful job as well (duh). Is there a way to mark a job successful? you can't continue a failed job. this is a known problem with unstable internet connections. maybe you can work around with openvpn (or something like that) to simply hide short outages to bacula. Another workaround could be splitting the fileset to several smaller ones. This way one job takes less time and it's more propable it will finish successfully. And if it won't, it takes less time to re-run it. -- TiN -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] continuing a failed job
Thomas Mueller wrote: Am Wed, 06 Jan 2010 11:08:17 +0200 schrieb Silver Salonen: Hi. Has anyone figured out how to continue a failed job? I have a client that has gigabytes of data, but very fragile internet connection, so it's almost impossible to get a normal full backup job. I thought I'd do it with VirtualFull job to create a successful job out of failed one and then use it as basis for incremental, but unfortunately VirtualFull requires a previous successful job as well (duh). Is there a way to mark a job successful? you can't continue a failed job. this is a known problem with unstable internet connections. maybe you can work around with openvpn (or something like that) to simply hide short outages to bacula. My first inclination is to say that if the network connection to the machine is sufficiently unstable that you can't complete a full backup, you probably shouldn't be trying to back it up over the network. However, there are ways to work around the problem. Were I trying to work around such a situation, with a mission-critical client on the far side of an unstable network connection, I would probably create a local partition somewhere of the same size as the disk on the remote machine, mirror the remote machine to that via rsync, and then back up the local mirror. The rsync will only need to transfer small amounts of data each day, and with the way rsync works, if one rsync is interrupted by a network outage, the next one will just pick up where it left off. Should you need to do a restore, you restore to the local mirror, then rsync only the restored files back to the remote client. -- 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, Free Stater It's not the years, it's the mileage. -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] continuing a failed job
On Wednesday 06 January 2010 19:30:42 Timo Neuvonen wrote: Thomas Mueller tho...@chaschperli.ch kirjoitti viestissä news:hi2ffj$ga...@ger.gmane.org... Am Wed, 06 Jan 2010 11:08:17 +0200 schrieb Silver Salonen: Hi. Has anyone figured out how to continue a failed job? I have a client that has gigabytes of data, but very fragile internet connection, so it's almost impossible to get a normal full backup job. I thought I'd do it with VirtualFull job to create a successful job out of failed one and then use it as basis for incremental, but unfortunately VirtualFull requires a previous successful job as well (duh). Is there a way to mark a job successful? you can't continue a failed job. this is a known problem with unstable internet connections. maybe you can work around with openvpn (or something like that) to simply hide short outages to bacula. Another workaround could be splitting the fileset to several smaller ones. This way one job takes less time and it's more propable it will finish successfully. And if it won't, it takes less time to re-run it. Well, yes.. I know the job itself cannot be continued. But could it be somehow marked as OK, so it could be used as a basis for another non-full backup? -- Silver -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users