(resending to list) On Tue, Feb 28, 2012 at 11:53 AM, Sławomir Skowron <slawomir.skowron@xxxxxxxxx> wrote: > > 2012/2/28 Yehuda Sadeh Weinraub <yehuda.sadeh@xxxxxxxxxxxxx>: > > On Tue, Feb 28, 2012 at 3:43 AM, Sławomir Skowron > > <slawomir.skowron@xxxxxxxxx> wrote: > >> After some parallel copy command via botto for many files everything, > >> going to slow down, and eventualy got timeout from nginx@radosgw. Note that you're overwriting the same object. Is that intentional? > > Reproduced logs in attachment. > I was able to recreate the issue. The problem is specifically related to the fact that you're overwriting the same object from 10s of parallel threads. What happens is that our race-detection code (that is related to the radosgw atomic object write) detects that the underlying object has been written and it needs to reread its header before overwriting it. This works well when there are a few writers to the same object, but doesn't scale very well. I opened issue #2120 to track the issue. Yehuda -- 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
