On 06/13/16 10:37 PM, Douglas K. Rand wrote: > On 06/13/16 14:00, Franck Ratier wrote: >> On Monday, 13 June 2016 10:10:24 UTC+1, [email protected] wrote: >>> I configured copies and it work fine. >>> >>> Now, I want perform some restorations tests but I have an issue : I >>> run "restore client=<client_name> copies". >>> >>> I show correctly jobID and copyjobId but it automatically select >>> jobId instead of copyjobId. If I continue (mark some files to >>> restore), it restore from original storage instead of storage of >>> copies. Looking at the implementation it only lists the copies indeed so it seems to be what the original design was when that was added. It does not update the list of JobIds its uses for restore. So I guess you just run into the fact that the design does not allow you to restore from copies.
>>> >>> But if I run restore with "restore client=<client_name> copies >>> jobid=<copyjobId>", it will restore from the storage of copies. >>> >>> If I specify "copies" in the first case, should it not use copy >>> storage directly ? That might be what you expect but that is not what the person who implemented it thought it should do or it was never finished. > >> I think that until the original job has not expired (i.e. retention >> time not expired), the copy "is just a copy". Once the original job >> expires the copy will be promoted and will be used for any restore >> you'd have to do. It's well explained in the docs, Chapter 22 - >> Migration and Copy. Correct normally you have the original Job on fast storage lets say disk and you copy to tape. So then you want to restore from disk and when things expire on disk the copy on take gets "promoted" to a normal Job and as such is available for restore. > > I find myself in exactly the same spot as Sylvain does: Testing that I > can restore from the tape based copies I made of my on-disk backups. > > I've tried several purge commands to try to remove the records of the > original backup from the catalog, but I still cannot restore from the > copy jobs. > > Any suggestions on how to go about this testing? Currently you probably need to migrate as then the original job gets removed. I might have a stab to see if a simple extra SQL query to retrieve the copies can be run when you say you explicitly want to use copies to restore. I could imagine that it makes sense when you loose a site and have sd-sd replication to an other datacenter. Problem is however that you can then only have one copy as otherwise it gets kind of difficult to select what copies to use. -- Marco van Wieringen [email protected] Bareos GmbH & Co. KG Phone: +49-221-63069389 http://www.bareos.com Sitz der Gesellschaft: Köln | Amtsgericht Köln: HRA 29646 Komplementär: Bareos Verwaltungs-GmbH Geschäftsführer: Stephan Dühr, M. Außendorf, J. Steffens, P. Storz, M. v. Wieringen -- 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.
