[Debian-ha-maintainers] Corosync not being stopped by init.d stop

Arturo Borrero Gonzalez cer.inet at linuxmail.org
Fri Aug 10 07:27:06 UTC 2012


Hi there!

Has anyone detected that in some some circumstances Corosync doesn't stop?

/etc/init.d/corosync stop|restart waits until infitine

Aug 10 09:21:16 cf02 crmd: [3839]: notice: do_state_transition: State
transition S_TRANSITION_ENGINE -> S_IDLE [ input=I_TE_SUCCESS
cause=C_FSA_INTERNAL origin=notify_crmd ]
Aug 10 09:21:16 cf02 crmd: [3839]: info: do_state_transition:
(Re)Issuing shutdown request now that we are the DC
Aug 10 09:21:16 cf02 crmd: [3839]: info: do_shutdown_req: Sending
shutdown request to cf02
Aug 10 09:21:16 cf02 crmd: [3839]: info: handle_shutdown_request:
Creating shutdown request for cf02 (state=S_IDLE)
Aug 10 09:21:46 cf02 corosync[3799]:   [pcmk  ] notice: pcmk_shutdown:
Still waiting for crmd (pid=3839, seq=6) to terminate...
Aug 10 09:21:48 cf02 clustersync[3755]: E: Rsync command failed while
working with node [cf01] Reason: ssh: connect to host cf01 port 22:
Connection timed out#015 rsync: connection unexpectedly closed (0
bytes received so far) [Receiver] rsync error: unexplained error (code
255) at io.c(605) [Receiver=3.0.9]
Aug 10 09:22:16 cf02 corosync[3799]:   [pcmk  ] notice: pcmk_shutdown:
Still waiting for crmd (pid=3839, seq=6) to terminate...
Aug 10 09:22:46 cf02 corosync[3799]:   [pcmk  ] notice: pcmk_shutdown:
Still waiting for crmd (pid=3839, seq=6) to terminate...
Aug 10 09:23:16 cf02 corosync[3799]:   [pcmk  ] notice: pcmk_shutdown:
Still waiting for crmd (pid=3839, seq=6) to terminate...
[...]

So I used to end with kill -9.

Maybe the init.d script should kill harder after a timeout?

I'm using: Corosync 1.4.2-3 in Debian Wheezy amd64


Best regards.
-- 
#
# Arturo Borrero Gonzalez || cer.inet at linuxmail.org
# Use debian gnu/linux!
#



More information about the Debian-ha-maintainers mailing list