[Linux-HA] Heartbeat 2 stops resources in wrong order

[Andrew Beekhof]

Hi Stephan,

can you include the complete CIB too please?
then I'll have enough to make a diagnosis pretty quickly.

Andrew

On 11/6/05, peinkofe at fhm.edu <peinkofe at fhm.edu> wrote:
> Hello everybody,
>
> while shutting down heartbeat I encountered, that it stops the resources in the wrong order.
> I use the current CVS HEAD version from 2005-11-05.
> What I did is:
> Stop heartbeat on the first node. Then it stopped the resources in the right order:
> Nov  6 13:52:07 sarek crmd: [9298]: info: mask(lrm.c:do_lrm_rsc_op): Performing op stop on telebase_rg:telebase_postmaste
> ..
> Nov  6 13:52:14 sarek crmd: [9298]: info: mask(lrm.c:do_lrm_rsc_op): Performing op stop on telebase_rg:telebase_data
> ..
> Nov  6 13:52:15 sarek crmd: [9298]: info: mask(lrm.c:do_lrm_rsc_op): Performing op stop on telebase_rg:telebase_ip
> ..
> Nov  6 13:52:16 sarek crmd: [9298]: info: mask(lrm.c:do_lrm_rsc_op): Performing op stop on telebase_rg:telebase_mailalarm
> ..
> Nov  6 13:52:17 sarek crmd: [9298]: info: mask(lrm.c:do_lrm_rsc_op): Performing op stop on telebase_rg:telebase_audioalarm
>
> Then waited some time until the second node took over the resources and shut down heartbeat at the second (last remaining) node.
> Nov  6 13:53:00 spock crmd: [21702]: info: mask(lrm.c:do_lrm_rsc_op): Performing op stop on telebase_rg:telebase_audioalarm
> ..
> Nov  6 13:53:01 spock crmd: [21702]: info: mask(lrm.c:do_lrm_rsc_op): Performing op stop on telebase_rg:telebase_ip
> ..
> Nov  6 13:53:02 spock crmd: [21702]: info: mask(lrm.c:do_lrm_rsc_op): Performing op stop on telebase_rg:telebase_mailalarm
> ..
> Nov  6 13:53:03 spock crmd: [21702]: info: mask(lrm.c:do_lrm_rsc_op): Performing op stop on telebase_rg:telebase_data
> ..
> Nov  6 13:53:04 spock crmd: [21702]: info: mask(lrm.c:do_lrm_rsc_op): Performing op stop on telebase_rg:telebase_postmaster
>
> This out of oder stop lead to the problem that one of my Filesystems was not unmounted, what ofcourse is potential dangerous if someone overlooks this and starts heartbeat up on the first node again.
>
> I attached the complete logfile of spock.
> Many thanks in advance.
> Stefan Peinkofer
>
>
>
> _______________________________________________
> Linux-HA mailing list
> Linux-HA at lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
>
>
>