Re: monitor 0.87.1 crashes

[samuel <samu60@xxxxxxxxx>]

apologies for the noise. Host  10.10.200.3 had some issues that made monitors to crash.

Thanks a lot for your help,

Samuel

On 27 March 2015 at 19:09, samuel <samu60@xxxxxxxxx> wrote:

Here it it goes (in case further information is needed, just ask and I would gladly offer it):

    -5> 2015-03-27 19:06:01.168361 7f94b4184700  5 mon.mon01@0(leader).osd e37404 send_incremental [37403..37404] to client.1419434 10.10.200.3:0/280

8592243

    -4> 2015-03-27 19:06:01.168427 7f94b4184700  1 -- 10.10.200.20:6789/0 --> client.1419434 10.10.200.3:0/2808592243 -- osd_map(37403..37404 src has

 36883..37404) v3 -- ?+0 0x3c59d40

    -3> 2015-03-27 19:06:01.168451 7f94b4184700  1 -- 10.10.200.20:6789/0 --> 10.10.200.3:0/2808592243 -- mon_subscribe_ack(300s) v1 -- ?+0 0x3fa4d00

 con 0x3c7e460

    -2> 2015-03-27 19:06:01.168465 7f94b4184700  1 -- 10.10.200.20:6789/0 <== client.1419434 10.10.200.3:0/2808592243 4 ==== pool_op(delete unmanaged

 snap pool 6 auid 0 tid 2617 name  v0) v4 ==== 65+0+0 (423335705 0 0) 0x3c33600 con 0x3c7e460

    -1> 2015-03-27 19:06:01.168475 7f94b4184700  5 mon.mon01@0(leader).paxos(paxos active c 16805455..16806016) is_readable = 1 - now=2015-03-27 19:0

6:01.168476 lease_expire=0.000000 has v0 lc 16806016

     0> 2015-03-27 19:06:01.170738 7f94b4184700 -1 ./include/interval_set.h: In function 'void interval_set<T>::insert(T, T) [with T = snapid_t]' thread 7f94b4184700 time 2015-03-27 19:06:01.168499

./include/interval_set.h: 340: FAILED assert(0)

 ceph version 0.87.1 (283c2e7cfa2457799f534744d7d549f83ea1335e)

 1: (ceph::__ceph_assert_fail(char const*, char const*, int, char const*)+0x85) [0x7d7825]

 2: /usr/bin/ceph-mon() [0x88def5]

 3: (pg_pool_t::remove_unmanaged_snap(snapid_t)+0x43) [0x886e53]

 4: (OSDMonitor::prepare_pool_op(MPoolOp*)+0xac5) [0x628d65]

 5: (OSDMonitor::prepare_update(PaxosServiceMessage*)+0x23b) [0x63b4eb]

 6: (PaxosService::dispatch(PaxosServiceMessage*)+0xd0f) [0x5f923f]

 7: (Monitor::dispatch(MonSession*, Message*, bool)+0x2a3) [0x5c0cf3]

 8: (Monitor::_ms_dispatch(Message*)+0x1cd) [0x5c178d]

 9: (Monitor::ms_dispatch(Message*)+0x23) [0x5e2443]

 10: (DispatchQueue::entry()+0x62a) [0x9194da]

 11: (DispatchQueue::DispatchThread::entry()+0xd) [0x7bc0cd]

 12: (()+0x7df3) [0x7f94bcfb0df3]

 13: (clone()+0x6d) [0x7f94bba931ad]

On 27 March 2015 at 19:04, Samuel Just <sjust@xxxxxxxxxx> wrote:

You'll want to at least include the backtrace.
-Sam

On 03/27/2015 10:55 AM, samuel wrote:

Hi all,

In a fully functional ceph installation today we suffer a problem with ceph monitors, that started crashing with following error:

include/interval_set.h: 340: FAILED assert(0)

Is there any related bug?

Thanks a lot in advance,

Samuel

_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com