Yes, that's true, you cannot migrate part of a job. If you migrated that job, then you would have unused space on some other volume :-(
__Martin >>>>> On Mon, 6 Sep 2021 11:30:20 +0200, Dan-Gabriel CALUGARU said: > > Thanks very much Martin for your answer. In fact, I had the both cases > to manage : > > > If the volume is not needed at all then maybe the "purge volume" command is > > what you need? That will forget about all data on a selected volume and > > mark > > it for reuse. > > Indeed, for the volume in this first case, I purged it and I think that > it will be no problem to fully reuse it later (for now, Bacula has taken > volumes marked Append). > > > > If the volume contains other jobs that you want to keep, then there is no > > way > > to reuse the space. The only way around that would be migrate the jobs to > > another volume. > > > For the volume in this second case, your response confirmed me that I > had to console myself with the idea of "losing" the 2 TB (out of 6 TB of > its total space). > > And this because I think that it is not possible to migrate only a part > of a job, I mean only the remaining 4 TB to another volume, as these 4 > TB are indeed useful, written by the previously successfully completed > job (this job consisted in ~25 TB written in 5 volumes: 3 voumes x6 TB > exclusively for this job, and 2 others partially written for this job, > including the one under discussion) > > > As a remark (who could perhaps make you better understand why I am > interested in these questions) : these backups are made for archiving > purposes (that's why I wanted to use as much as possible of the volumes > capacity and so ideally not to have "lost" spaces, because if they will > be some ones, they will be so for a long time...) > > > Have a good week, > > Dan > > > > __Martin > > > > > >>>>>> On Fri, 3 Sep 2021 14:13:34 +0200, Dan-Gabriel CALUGARU said: > >> Thank you Josh and Radosław for your additional responses. > >> > >> But in this case (for a Failed job due to bacula machine crash), restart > >> is the same as (re)run the job, right ? > >> > >> And thus, I come back to my main question (from my initial post): how to > >> properly free up the space written by the Failed job (more than 3TB) on > >> the volumes in order to be reused ? > >> As a reminder, during this writing : > >> - the first used volume became Full (it already contained data correctly > >> written by the previous successfully completed job) > >> - and continued writing on another volume, before the machine crashed > >> (this volume is now used for the restart job) > >> > >> This is why (in a previous try) I did my tests with the purge command > >> for this jobID (which presumably hadn't purged anything), and delete > >> command this jobID which worked to remove this jobID informations in the > >> catalog. But not to free up space written by this jobID on volumes. > >> > >> In this case, what is the theoretical difference between purge and delete ? > >> > >> I also wanted to use the Action On Purge = Truncate feature for both > >> volumes before initiating the purge of this jobID, but since purge > >> doesn't seem to do anything for this jobID, then I'm instead wondering > >> if this feature (or another equivalent) would also work for the delete > >> command (which it seemed to work in my previous try)? > >> > >> Sorry if my questions are too basic, but I haven't really found a clear > >> answer in the manuals for these points. > >> > >> Have a good day, > >> > >> Best regards > >> > >> Dan > >> > >> > >> Le 03/09/2021 à 13:13, Radosław Korzeniewski a écrit : > >>> Hello, > >>> > >>> pt., 3 wrz 2021 o 12:36 Josh Fisher <jfis...@jaybus.com > >>> <mailto:jfis...@jaybus.com>> napisał(a): > >>> > >>> Instead, the Bacula server machine itself crashed, so it cannot > >>> determine where to restart and did not retain any cached/spooled > >>> data. > >>> > >>> Yes, exactly. I did miss this very important information in the > >>> original post. It changes everything! > >>> > >>> But the other statements about (not) purging files and job info are > >>> still valid. > >>> > >>> best regards > >>> -- > >>> Radosław Korzeniewski > >>> rados...@korzeniewski.net <mailto:rados...@korzeniewski.net> > >>> > >>> > >>> _______________________________________________ > >>> Bacula-users mailing list > >>> Bacula-users@lists.sourceforge.net > >>> https://lists.sourceforge.net/lists/listinfo/bacula-users > > -- > Dan-Gabriel CALUGARU > IR en Calcul Scientifique (CNRS) > Dr de Mathématiques et Applications > > Laboratoire de Mécanique des Fluides et d'Acoustique > UMR 5509 CNRS - ECL - UCBL - INSA Lyon - Univ. de Lyon > > Bâtiment I11 - bureau 11098 > ECOLE CENTRALE de LYON > 36, avenue Guy de Collongue > 69134 ECULLY > > tel: +33 (0)4 72 18 61 73 > > _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
