Re: MDS crash, wont startup again

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, May 21, 2012 at 5:38 AM, Felix Feinhals
<ff@xxxxxxxxxxxxxxxxxxxxxxx> wrote:
> Hi Josh,
>
> i quoted the trace and some other stats in my first email, maybe it
> got stuck in the spam filters.
> Well next try:
>
> snip
>
> -3> 2012-05-10 14:52:29.509940 7fb1c9351700 1 mds.0.40 handle_mds_map
>  i am now mds.0.40
>  -2> 2012-05-10 14:52:29.509956 7fb1c9351700 1 mds.0.40 handle_mds_map
>  state change up:reconnect --> up:rejoin
>  -1> 2012-05-10 14:52:29.509963 7fb1c9351700 1 mds.0.40 rejoin_joint_start
>  0> 2012-05-10 14:52:29.512503 7fb1c9351700 -1 *** Caught signal
>  (Segmentation fault) **
>  in thread 7fb1c9351700
>
> ceph version 0.46 (commit:cb7f1c9c7520848b0899b26440ac34a8acea58d1)
>  1: ceph-mds() [0x814279]
>  2: (()+0xeff0) [0x7fb1cddbfff0]
>  3: (SnapRealm::have_past_parents_open(snapid_t, snapid_t)+0x4f) [0x6cb5ef]
>  4: (MDCache::check_realm_past_parents(SnapRealm*)+0x2b) [0x55d58b]
>  5: (MDCache::choose_lock_states_and_reconnect_caps()+0x29c) [0x572eec]
>  6: (MDCache::rejoin_gather_finish()+0x90) [0x5931a0]
>  7: (MDCache::rejoin_send_rejoins()+0x2c05) [0x59b9d5]
>  8: (MDS::rejoin_joint_start()+0x131) [0x4a8721]
>  9: (MDS::handle_mds_map(MMDSMap*)+0x2c4a) [0x4c253a]
>  10: (MDS::handle_core_message(Message*)+0x913) [0x4c4513]
>  11: (MDS::_dispatch(Message*)+0x2f) [0x4c45ef]
>  12: (MDS::ms_dispatch(Message*)+0x1fb) [0x4c628b]
>  13: (SimpleMessenger::dispatch_entry()+0x979) [0x7acb49]
>  14: (SimpleMessenger::DispatchThread::entry()+0xd) [0x7336ed]
>  15: (()+0x68ca) [0x7fb1cddb78ca]
>  16: (clone()+0x6d) [0x7fb1cc63f92d]

There's nothing obvious here — can you run gdb on the core and get
another backtrace and the info from levels 3-5?

> snip
>
> I though ceph chooses which MDS is active and which is standby, i just
> have 3 in the cluster config:
Yes, it does — if you don't increase the number of allowed MDSes
you'll just get one of them active.
-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


[Index of Archives]     [CEPH Users]     [Ceph Large]     [Information on CEPH]     [Linux BTRFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]
  Powered by Linux