Kern Sibbald wrote: > My best guess is that the indexes did not get properly > generated some where > along the line. Whether the indexes were not there in 1.36.1 > or they got > lost or need rebuilding when upgrading I am not sure -- this > seems to have > happened to other users. > > I recommend that you read the MySQL chapter in the current > online manual. It > tells you which indexes you should have, and in the make_mysql_tables > script, there are some comments which suggest additional indexes you > may want to add depending on your site setup. > > MySQL is relatively fault tolerant, so if you find that you > are missing > indexes, I would be inclined to stop the Director, build the > appropriate ones, and try again. > > Also take a look at your my.ini file -- depending on your > system size, you > might want to modify it (I think the manual has something on this). > > On Monday 19 February 2007 17:44, Andrew Paterson wrote: >> Hi All, >> Solaris 9 (sun4u) Bacula-1.36.1 working (worked!) fine -> upgrading >> to bacula-2.0.2. [Mysql 4.0.21 Solaris 8,9,10, Linux 2.6 (FC5 & >> GENTOO), WIN-32 (WIN-XP) Clients.] >> >> (Bacula built from source). >> >> Upgrading initially just the backup server - >> comprising director on a dedicated Solaris 9 box with bacula-sd >> (driving a storedge (HP) autoloader) and Mysql DB (& of course >> bacula-client & bconsole). >> >> [database version when running 1.36.1 was 8] >> Ran updatedb/update_mysql_tables_8_to_9 .................. - ok >> Ran /etc/bacula/update_mysql_tables [updated tables to V10] - ok >> >> Bacula-dir starts OK (initial problem with "Accept Any Volume = yes" >> in Pool directive - commented it out). >> >> bconsole starts & connects - ok >> - "status stor" - works fine. >> - But when I do "status dir" >> bconsole gets as far as >> %<--------- >> Scheduled Jobs: >> Level Type Pri Scheduled Name >> Volume >> > ============================================================== > ============= >> ===== %<---------- >> And appears to hang. (4 hrs so far!) >> >> mysql administrator shows DB server running a query "DELETE FROM >> File WHERE JobId=9032" trouble is it shows "processing time" as the >> same all the time!!. Actually, to be honset JobId has changed at >> last so maybe its just a little slow. Why should it take so long? Is >> it blowing away my backup database (which I have to admit is pretty >> big!). Maybe I am just being impatient (neurotic) since usually the >> DB pruning is done early in the morning whilst I am blissfuly >> unconcious! However, when I run the director with debugging on >> (-d10) the log (/var/bacula/working/log in my case) shows absolutely >> nothing additional. Where does the tracing go? >> >> However, what throws me the most is that since the director is >> multi-threading, does the channel to bconsole have o be locked out >> because the director is doing a purge database operation? how about >> a "busy" response then?. >> >> Any suggestions/clue requests/enlightenments to calm my jangling >> nerves ????? >> >> Andrew R Paterson >> DS Ltd. >> Cedar Office Park >> Cobham Road >> Ferndown >> Dorset BH21 7SB >> >> Tel: +44 (0) 1202 877188 >> Fax: +44 (0) 1202 871297 >> www.ds.co.uk >> >> > -------------------------------------------------------------- > ----------- >> 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 Hi Kern et al, thanks very much for that diagnosis. That sounds very much like the cause - I will investigate my indeces! Happily, on leaving the director running overnight and returning this morning - it has completed its pruning. It appears I have upgraded just prior to the pruning of 4 volumes from the catalogue!. (328 jobs, 523 jobs, 225 jobs & 482 jobs respectively). Either that or the new director has a different view of when it should prune. I think I need to review my bacula-dir.conf regarding Retention periods - something strange is going on. On the good side, (predictably) I still have a good (better?) bacula database and having left the new director running overnight, it naturally started up the overnight backups (but about half were aborted due to max start-time exceeded dus to the DB prune). I'm lucky that didn't cause more problems because the clients are still running 1.36.1 bacula-fd's but they seemed happy. I shall upgrade the clients today. Actually, it turns out that the upgrade went pretty well considering the way I just "went for it!". Any problems that occurred were totally my fault and highlight a major problem with something like bacula..... One learns all about it, builds & installs it, lets it run & run & run & forget about it!!!!!!! Then one dark night something happens or you need to upgrade - but all you learnt last year has been "pruned" from your brain :). Still being picky tho' - can't something be done about a busy director blocking bconsole? Many many thanks Kern for an excellent backup system Regards Andrew R Paterson DS Ltd. Cedar Office Park Cobham Road Ferndown Dorset BH21 7SB Tel: +44 (0) 1202 877188 Fax: +44 (0) 1202 871297 www.ds.co.uk
------------------------------------------------------------------------- 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
