Hello, pt., 15 mar 2019 o 10:55 Thorsten Johannsen <thorsten-johann...@gmx.net> napisał(a):
> Hi, > > On 14.03.19 22:12, Radosław Korzeniewski wrote: > > Hello, > > > czw., 14 mar 2019 o 16:05 Thorsten Johannsen <thorsten-johann...@gmx.net> > napisał(a): > >> (...) inside volume for jobID 5269 (...) > > > There is no such functionality in Bacula as described above. You cannot > reserve or dedicate a volume for particular jobid. > > All you can do is a dedicate a pool of volumes for a Job. > > Oh dear, then my whole concept just blew up. > Sorry about it... > But can you point me at my logical error? > > In my mind, if > > - Volume autolabel is enabled > > - Maximum jobs per volume is set to 1 > > - Recycling of volumes is disabled > > > Bacula would be forced to create (label) a new volume for each job > Wrong! Bacula would label a new volume if no volumes will be available in the Pool only. > , wouldn't it? Where am I wrong here? > The label procedure and selecting a volume for backup are the two different operations. Just take a look at the following algorithm: 1. select a required Pool for a particular Job 2. search for any (it is not exactly true) volume with Append or Recycle status and use it for backup (sort of) 3. if not, make a recycle procedure in this Pool and search for volumes with Purged status and use it for backup (sort of, as always) 4. if not, check for Scratch Pool parameter or just a pool with name Scratch and get a first (it is not exactly true) volume with matched Media Type (sort of, as always) 5. if not then check if the automatic volume labeling is enabled for this Pool and Device, if yes then create the new volume (it will have an Append status) in this Pool using a selected Device. FORGET everything you've done until now and GOTO 1. :) There is a p.6. which inform an operator about backup problem, but it is not a case here. The above procedure has a lot of special cases when it goes totally different because it is configured so, but it will blur the overview. So, as you can see the only thing you can have guaranteed is a Job stored in the selected Pool, not a volume. The above algorithm is executed by every Job, every time it requires a volume. But at the end of the day it really doesn't matter what volume was used for backup, because all required tracking information is stored in catalog database for you. I hope it helps. -- Radosław Korzeniewski rados...@korzeniewski.net
_______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
