Hi Thomas, yes, I did that already and set this timer to 120 minutes. But this was not enough obviously or the scheduled backup time was to early. I changed the backup time now to start later at 4:00 and now the tests are ready before backup and the backup is ready before the nightly build at 5:30. Works perfect again.
The "problem" occured because I am now generating those test projects programmatically via groovy and because of that we are now getting much better test coverage (with longer run times of course)... Thanks for listening.;-) Dirk 2012/7/4 Thomas Fürer <[email protected]>: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Hi Dirk, > > Thank you for using Thinbackup. > There is a version out where you can choose your own time when the > backup is force a quiet mode. if you increase the time jobs in the > queue have a chance to be executed until, but if the system gets idle > earlier you do not waste time and the backup starts. > > thomas > > Am 04.07.2012 17:40, schrieb Dirk Kuypers: >> I think I found it... >> >> I am using ThinBackup plugin for incremental backups during night >> and it had a check to force Jenkins into quiet mode after 120 >> minutes of being not idle. >> >> Hm, I think I have to reschedule some things... >> >> BR Dirk >> >> 2012/7/4 Dirk Kuypers <[email protected]>: >>> Hi list, >>> >>> I am just puzzling about some strange behaviour of my >>> master/slave setup. I have a 4 core master with Win7/64 bit and >>> about 10 slaves with 2 cores and Win7/64 bit/Win2008 Server R2. I >>> compile a .NET project at 20:00 and start (at the moment as a >>> scheduled job at 21:00) about 150 unit test projects to keep the >>> cluster chewing during night. If you look at the attached picture >>> you can see that there are about 120 queued jobs when the test >>> projects start which get less over time. Somewhere around 2:30 >>> the number of occupied executors begins to drop and no queued job >>> gets executed any more. At 5:30 the regular nightly compile and >>> setup generation of our software is started, but it is only >>> queued and there is _no_ dependency to the unit test jobs (block >>> build because up/downstream job is running). Suddenly at around >>> 7:00 the queued jobs begin to be executed again. >>> >>> There is nothing conspicuous in the logs of the jobs. Any ideas >>> if this is a known bug or where I could look for direct Jenkins >>> logs telling me why a job is waiting and not executed? >>> >>> Thanks for hints Dirk >> > > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.4.10 (MingW32) > Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ > > iQEcBAEBAgAGBQJP9GlXAAoJEMIrIJt+IWiiZEIH/iGpXeouVIydSypYgUWHjzUq > ZACmk416m2TitIVtmdYGzaMj/IchP4TyWmL8HGED5hO7E0CcJEh0Apk/sXdnQoMS > ENiNF/EfebC+cIkX2+AyV7PBEVegdADwiSCs/p5qvUbQUFgCNSfnkM6Db2Ceo60O > TSFHzjHH+zwSovUzNCE4GhpPbSbKf6i/TNUbz0XXMjayIGTOVuE3yLFXwdWh4w3E > qLFEXDF5SP8W3u+3m57q+rAQGLz/GjyxBJ98McEgSISCzbpFw9w8B12lZWpEML/P > J9Ydqic/+PgVHW1CqiSMPYV6idihuTUsPSkSqu8Y1MORBu63G8terEQ0JPoQB90= > =o8DC > -----END PGP SIGNATURE-----
