On Sat, 23 Feb 2013, Chris Dunlop wrote: > On Fri, Feb 22, 2013 at 03:43:22PM -0800, Sage Weil wrote: > > On Sat, 23 Feb 2013, Chris Dunlop wrote: > >> On Fri, Feb 22, 2013 at 01:57:32PM -0800, Sage Weil wrote: > >>> On Fri, 22 Feb 2013, Chris Dunlop wrote: > >>>> G'day, > >>>> > >>>> It seems there might be two issues here: the first being the delayed > >>>> receipt of echo replies causing an seemingly otherwise healthy osd to be > >>>> marked down, the second being the lack of recovery once the downed osd is > >>>> recognised as up again. > >>>> > >>>> Is it worth my opening tracker reports for this, just so it doesn't get > >>>> lost? > >>> > >>> I just looked at the logs. I can't tell what happend to cause that 10 > >>> second delay.. strangely, messages were passing from 0 -> 1, but nothing > >>> came back from 1 -> 0 (although 1 was queuing, if not sending, them). > > Is there any way of telling where they were delayed, i.e. in the 1's output > queue or 0's input queue? Yeah, if you bump it up to 'debug ms = 20'. Be aware that that will generate a lot of logging, though. > >>> The strange bit is that after this, you get those indefinite hangs. From > >>> the logs it looks like the OSD rebound to an old port that was previously > >>> open from osd.0.. probably from way back. Do you have logs going further > >>> back than what you posted? Also, do you have osdmaps, say, 750 and > >>> onward? It looks like there is a bug in the connection handling code > >>> (that is unrelated to the delay above). > >> > >> Currently uploading logs starting midnight to dropbox, will send > >> links when when they're up. > >> > >> How would I retrieve the interesting osdmaps? > > > > They are in the monitor data directory, in the osdmap_full dir. > > Logs from midnight onwards and osdmaps are in this folder: > > https://www.dropbox.com/sh/7nq7gr2u2deorcu/Nvw3FFGiy2 > > ceph-mon.b2.log.bz2 > ceph-mon.b4.log.bz2 > ceph-mon.b5.log.bz2 > ceph-osd.0.log.bz2 > ceph-osd.1.log.bz2 (still uploading as I type) > osdmaps.zip I'll take a look... sage -- 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
