On Tuesday 10 October 2006 07:21, Michael Brennen wrote: > > Kern, this is a brief update to the list but primarily with you in > mind. > > Curiously, as of cvs a few days ago the system is now using multiple > tape drives again, without any change at all in the configuration. > For the last few days bacula is not forcing multiple jobs destined > to different volumes to all go through tape drive 0. Since I always > note these things after the fact it is difficult to correlate with > certainty to a given update, but it *might* have been the 1.39.24 > cvs. I updated cvs today to 1.39.25. > > I am still looking at the regression tests (against current bacula > 1.39.25 cvs) and trying to understand why a particular tape load is > failing but the regression test sometimes goes on to run > successfully anyway. Understanding that may well take some time, as > testing is interleaved with everything else going on, but I am > trying to be respectful of your time.
Yes, what you write doesn't surprise me. I may be wrong, but I cannot remember *any* changes that would affect what you are seeing, and that for quite a long time now. I think you are seeing not a bug but a facet of Bacula that is not fully developed. That is if a drive is not already busy, it will make use of the first one it comes across. This may occur even though the tape it will eventually need is on a different drive. Once it locks on to a drive, it will not switch drives (this will come in a later version), so in some cases, Bacula needs to unload one drive then load the volume it just unloaded into another drive. This is not the most efficient, but it works. The above is all very "timing" dependent. By that I mean what job is running when, and the timing is highly influenced by how much data is being backed up. If you are running multiple simultaneous jobs and the first job finishes quickly the second job will use the same job giving the appearance that it only uses one drive, but in fact for Bacula, it finds no drives busy so just selects one, which may often be the same one. Anyway, bottom line: after a lot of testing here, I think Bacula is functioning within the design limitations (mainly once it has a drive it will not switch), and what you were seeing was not a bug but an ineffiency that will be corrected by a future enhancement. Concerning your autochanger problems with the Signal 15, I am convinced this is a timout problem probably due to a rewind that took a long time. Regards, Kern > > Thanks again... > > -- Michael > > ------------------------------------------------------------------------- > Take Surveys. Earn Cash. Influence the Future of IT > Join SourceForge.net's Techsay panel and you'll get the chance to share your > opinions on IT & business topics through brief surveys -- and earn cash > http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV > _______________________________________________ > Bacula-users mailing list > Bacula-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/bacula-users > ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
