On 09/07/16 05:38, Philipp Storz wrote: > Hello Douglas, > > On 02.09.2016 17:27, Douglas K. Rand wrote: >> I've been playing around with using Virtual Full backups for off-site >> backups, and generally I really like it. But of course I have a problem. >> >> First, a bit of background: We are doing disk to disk backups for daily >> backups. I have the full backups spread across the week to even the load >> and keep the backup window sized about the same across the week. >> >> Our off-site backups happen once a week and go onto LTO-6 tapes and then >> are stored, well, not on-site. :) We keep a few months of tapes stored >> off-site and then eventually re-use the tapes. >> >> I was doing Copy jobs to make the off-site tapes, but then I had to >> schedule full backups close to the Copy job for them to be fairly fresh. >> Also, since you can't recover from a Copy job until the source job has >> been purged, testing the off-site backups was a pain. >> >> So I tried Virtual Full and they are just as fast as Copy jobs for me. >> (Even faster if I allow the Virtual Full job to spool to disk first.) >> And I get the freshest data for the backups. >> >> But the problem is that sometimes, especially during my testing, the >> Virtual Full tries to use the last Virtual Full backup as the base. >> Which doesn't work since that is on tape and not on disk. >> >> Is there a way to get the Virtual Full job to ignore the backups stored >> on tape when it does a new one, and only use the on-disk Full backups as >> the base? All of them are in different pools, so I had thought that that >> would do it, but it doesn't. > > Yes, there is a way which is admittedly not really nice yet. > > I have already an idea how to do this cleanly, but for now it > should suffice to your requirements. > > Simply change the job type of your virtual full from "backup" to "archive", > for > now you need to do this in the database manally. > > In Master you can already do "update jobid=xxx set jobtype=A" > > This will make sure that it is not taken as base for following backups.
Thanks for the response Phillip. Would the SQL command be more like: update job set type='A' where jobid=xxx; Where, of course, xxx is the job id of the Virtual Full backup. It doesn't seem that the bconsole update command allows what you suggested. (Unless I'm way off base.) Does setting the jobtype to A(rchive) affect restores at all? Will I be able to restore from an Archive job just as I restore from a Backup job? Regards, Doug. -- You received this message because you are subscribed to the Google Groups "bareos-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. For more options, visit https://groups.google.com/d/optout.
