On Thu, May 28, 2015 at 2:32 AM, Loic Dachary <loic@xxxxxxxxxxx> wrote: > Hi, > > This morning I'll schedule a job with priority 50, assuming nobody will get mad at me for using such a low priority because the associated bug fix blocks the release of v0.94.2 (http://tracker.ceph.com/issues/11546) and also assuming noone uses a priority lower than 100 just to get in front of the nightlies[1]. In my imagination > > priority [0,100] is for emergencies > priority [100,1000] is to schedule a job with higher priority than the nightlies > priority 1000 (the default) is for all automated tests and no human being wait on them (the nightlies for instance). > > Does someone have a different mapping in mind ? Those are pretty sensible. In practice I think the important points to be aware of are: 1000: nightly priority 101: Sam's priority for anything he schedules. That 101 can be a bit tricky since going over it can mean you don't run for a few days if he's busy running stuff repeatedly. It's kind of an unfortunate race to the bottom when there are several people who think their stuff is important. (A few weeks ago Sage was running RGW bugfix tests for several days, at least some of which were priority 1. I was also trying to push some stuff through at the time and that was quite an experience.) :/ -Greg -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
