The file .bsr that generate restore (bconsole) Can i used for bextrack? for a restore very fast
On Wed, May 18, 2022 at 9:07 AM Jose Alberto <j.se...@gmail.com> wrote: > I think that is where we failed. We don't have bsr file. > > Our bacula is a default installation from the official Bacula repositories. > > I assume that I must enable so that each Job generates a bsr file. > > With the bsr do I avoid the bscan? Would you only use bextrack? > > On Wed, May 18, 2022 at 5:48 AM Radosław Korzeniewski < > rados...@korzeniewski.net> wrote: > >> Hello, >> >> wt., 17 maj 2022 o 23:45 Jose Alberto <j.se...@gmail.com> napisał(a): >> >>> They requested a backup (datapump) of 1.5 TB but from Dec 15, 2021. The >>> Job for fileretention had already expired. >>> >>> The volume was as FULL and with 25 TB of data. >>> >>> 1) I tried first with bextrack and it's too slow. 6 hours without any >>> file. >>> >> >> With a bootstrap file you should be able to restore all what you need >> very fast. >> How did you execute bextract, what parameters? >> >> >>> >>> 2) try using bconsole restore Regexp. I end up with status OK after 10 >>> hours, but with 0 files recovered. >>> >>> 3) Well, I filled myself with patience and went to search "bscan". Which >>> lasted 40 hours to complete the recreation of the catalog. And I managed to >>> restore the requested 1.5 TB file in just 4 hours. >>> >>> After this desperate experience but with the expected results. The >>> following questions arise. >>> >>> 1) Is there a way to make the bscan run faster? I had always worked with >>> lto6. But lto7 is already different, it's 25TB of data with a FULL tape, I >>> assume that the search time increases. >>> >> >> With proper bootstrap files you will not need to use bscan for this. >> >> >>> >>> 2) Is there any other way to create the Files in the catalog but only >>> indicating the Jobid? >>> >> >> Well, what you did is a full metadata scan of your archive which >> populated the catalog database and during restore the director created a >> bootstrap file for you and did restore using this file. So if you will have >> a proper bootstrap file from start then all data scanning won't be required. >> >> >>> >>> Just imagine a BSCAN in lto8 or 9 >>> >>> 3) What can be the best path for these scenarios? >>> >>> >> Archive all your bootstrap files in any case you will need to restore >> some data from your tapes. >> >> >>> Due to this experience we increased the times in fileretention and >>> jobretention in the most critical servers. >>> >> >> +1 >> >> Now, when database storage is huge and very fast, there is no real world >> reason to lower file and job retention below volume retention. >> And always archive all generated bootstrap files. They will save a lot of >> your time. >> >> best regards >> -- >> Radosław Korzeniewski >> rados...@korzeniewski.net >> > > > -- > ############################# > # Sistema Operativo: Debian # > # Caracas, Venezuela # > ############################# > -- ############################# # Sistema Operativo: Debian # # Caracas, Venezuela # #############################
_______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
