>>> Lars Ellenberg <[email protected]> schrieb am 25.11.2011 um 13:29 in Nachricht <[email protected]>: > On Fri, Nov 25, 2011 at 11:34:33AM +0100, Florian Haas wrote: > > On 11/25/11 10:47, Ulrich Windl wrote: > > > The resource is unmanaged: > > > Nov 24 12:59:05 h03 pengine: [15876]: notice: LogActions: Leave > prm_c11_db_15k_raid1 (Started unmanaged) > > > [...] > > > LUN shrink begins: > > > Nov 24 12:59:39 h03 kernel: [1220873.890571] sd 2:0:3:13: [sdai] Result: > hostbyte=DID_OK driverbyte=DRIVER_SENSE > > > [...] > > > RAID monitor runs: > > > Nov 24 13:08:31 h03 lrmd: [15874]: info: rsc:prm_c11_db_15k_raid1 > monitor[262] (pid 28046) > > > Nov 24 13:08:31 h03 Raid1[28046]: [28061]: WARNING: /dev/md10 has at > > > least > one failed device. > > > Nov 24 13:08:31 h03 Raid1[28046]: [28063]: INFO: Attempting recovery > sequence to re-add devices on /dev/md10: > > > Nov 24 13:08:31 h03 lrmd: [15874]: info: RA output: > (prm_c11_db_15k_raid1:monitor:stderr) mdadm: re-added > /dev/disk/by-id/dm-name-C11_DB_15k-E2 > > > Nov 24 13:08:31 h03 kernel: [1221405.912376] md: bind<dm-48> > > > Nov 24 13:08:31 h03 kernel: [1221405.913568] RAID1 conf printout: > > > Nov 24 13:08:31 h03 kernel: [1221405.913572] --- wd:1 rd:2 > > > Nov 24 13:08:31 h03 kernel: [1221405.913576] disk 0, wo:0, o:1, dev:dm-14 > > > Nov 24 13:08:31 h03 kernel: [1221405.913579] disk 1, wo:1, o:1, dev:dm-48 > > > Nov 24 13:08:31 h03 kernel: [1221405.913706] md: recovery of RAID array > md10 > > > [...] > > > (The resource still is unmanaged) > > > Nov 24 13:14:06 h03 pengine: [15876]: notice: LogActions: Leave > prm_c11_db_15k_raid1 (Started unmanaged) > > > > > > So you see. > > > > I agree that this clearly ought not to happen. > > > > From the log snippet it's > > not entirely clear whether that's a recurring monitor (interval == > > whatever you configured, or 20 if default), or a probe (interval == 0). > > > > A recurring monitor clearly should not happen at all when unmanaged. > > That is incorrect. > > is-managed=false does still monitor the resource. It only prevents > pacemaker from sending start/stop etc commands to that resource. > > If the implementation of the monitor action in the RA does trigger > "auto-recovery" or other things, well, then it does.
Well, IMHO the RA should not try auto-recovery in "unmanaged" mode. I'm unsure whether I can set "maintenance-mode" for a single resource. Regards, Ulrich > > If you don't want that, you'd need to either go "maintenance-mode=true", > (which I'd recommend; it simply does no actioin at all). > Or disable monitor operations on the is-managed=false resource as well. > > > A > > probe, however, would cause the RA to skip this part of monitor. And, it > > would be skipped altogether if OCF_CHECK_LEVEL == 0. > > In this case it would be sufficient to disable only the monitor > action(s) with OCF_CHECK_LEVEL > 0. _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
