Activity
From 04/18/2016 to 05/17/2016
05/17/2016
05/16/2016
- 04:22 PM Feature #5914 (Rejected): msgr: use either socket or rsocket calls
- 04:22 PM Feature #5913 (Rejected): msgr: add wrappers for remaining socket calls in Pipe, Accepter
- 04:22 PM Feature #5912 (Rejected): msgr: create rsockets entity_addr_t type
05/12/2016
- 04:09 AM Bug #15849: msg/async: connection race hang
- https://github.com/ceph/ceph/pull/9086
- 02:42 AM Bug #15849: msg/async: connection race hang
- it looks something wrong with nonblock connect logic, I would like to see the whole log file so I can find out the re...
05/11/2016
- 05:18 PM Bug #15849: msg/async: connection race hang
- 05:16 PM Bug #15849 (Resolved): msg/async: connection race hang
- osd.301 and osd.344 are tryign to talk to each other.
osd.301 log includes (this is debug ms = 10)...
05/10/2016
05/07/2016
- 02:34 AM Bug #15758: msgr/async: Messenger thread long time lock hold risk
- yes, async messenger always easy to expose problem than simple. actually this pr(https://github.com/ceph/ceph/pull/88...
05/06/2016
- 07:01 PM Bug #15758: msgr/async: Messenger thread long time lock hold risk
- Is this something that newly blocks for a long time?
Or is the problem that AsyncMessenger doesn't have per-connec... - 07:48 AM Bug #15758: msgr/async: Messenger thread long time lock hold risk
- It looks like async messenger need a extra DispatchQueue to handle nonfast message which may block thread totally.
- 07:08 AM Bug #15758: msgr/async: Messenger thread long time lock hold risk
- The lock time depends on the ObjectStore process time from 1s-5s even longer. It's easy to cause IO shake when exists...
- 07:03 AM Bug #15758 (Resolved): msgr/async: Messenger thread long time lock hold risk
- Thread 46 (Thread 0x7fa7365a2700 (LWP 17842)):
#0 0x00007fa73b888705 in pthread_cond_wait@@GLIBC_2.3.2 () from /lib...
05/05/2016
- 04:37 PM Bug #14120 (Resolved): Pipe::do_recv() may loop infinitely
- 12:35 PM Bug #14120: Pipe::do_recv() may loop infinitely
- I encountered this again. I ran 'perf record -g -e cycles:u -p xxx', perf report > perf.out.txt
- 01:24 PM Bug #15503 (Rejected): msg/async: deadlock in rebind when enabling delay
04/28/2016
04/27/2016
- 08:01 PM Bug #15517: msg/simple: PipeConnection leak
- Could the OSD have lost this, rather than the Messenger?
Also available in: Atom