[sheepdog-users] simulating network problems with a 3 nodes cluster

[Valerio Pachera]

Unplug power from the switch for 5 seconds.

All 3 sheeps daemons die.

Sheep.log

3 nodes, shutdown switch ~30''

Nov 15 15:50:33  EMERG [main] cdrv_cpg_confchg(573) PANIC: Network
partition is detected
Nov 15 15:50:33  EMERG [main] crash_handler(254) sheep exits unexpectedly
(Aborted).
Nov 15 15:50:33  EMERG [main] sd_backtrace(805) sheep.c:256: crash_handler
Nov 15 15:50:33  EMERG [main] sd_backtrace(819)
/lib/x86_64-linux-gnu/libpthread.so.0(+0xf02f) [0x7f9426d8002f]
Nov 15 15:50:33  EMERG [main] sd_backtrace(819)
/lib/x86_64-linux-gnu/libc.so.6(gsignal+0x34) [0x7f942638c474]
Nov 15 15:50:33  EMERG [main] sd_backtrace(819)
/lib/x86_64-linux-gnu/libc.so.6(abort+0x17f) [0x7f942638f6ef]
Nov 15 15:50:33  EMERG [main] sd_backtrace(805) corosync.c:573:
cdrv_cpg_confchg
Nov 15 15:50:33  EMERG [main] sd_backtrace(819)
/usr/lib/libcpg.so.4(cpg_dispatch+0x594) [0x7f94268ebd74]
Nov 15 15:50:33  EMERG [main] sd_backtrace(805) corosync.c:703:
corosync_handler
Nov 15 15:50:33  EMERG [main] sd_backtrace(805) event.c:210: do_event_loop
Nov 15 15:50:33  EMERG [main] sd_backtrace(805) sheep.c:843: main
Nov 15 15:50:33  EMERG [main] sd_backtrace(819)
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xfc) [0x7f9426378eac]
Nov 15 15:50:33  EMERG [main] sd_backtrace(819) sheep() [0x405698]
Nov 15 15:50:33  ERROR [main] crash_handler(528) sheep pid 5931 exited
unexpectedly.


With 2 nodes only, the deamons didn't die.
I'll try to repeat the the test later with 2 node and see if till behave
the same way.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.wpkg.org/pipermail/sheepdog-users/attachments/20131115/45cab553/attachment-0004.html>