[Pacemaker] failure of a monitor operation would be disappeared from crm_mon

Andrew Beekhof beekhof at gmail.com
Mon Sep 8 07:14:31 EDT 2008


On Sep 8, 2008, at 10:11 AM, Junko IKEDA wrote:

> Hi,
>
> This is a requet about showing monitor NG with crm_mon.
> One dummy resource is running like this;
>
> # crm_mon -fot -i1 -r
> Node: node-b (59295d90-5459-490d-a1e0-d48810cf2fb3): online
> Node: node-a (b3852a23-c10b-440a-a8e0-263b0185d657): online
>
> dummy   (ocf::heartbeat:Dummy): Started node-a
>
> Operations:
> * Node node-b:
> * Node node-a:
>   dummy:
>    + start: rc=0 (ok)
>    + monitor: interval=10000ms rc=0 (ok)
>
>
>
> remove its status file, so dummy will do failover.
>
> # rm -f /var/run/heartbeat/rsctmp/Dummy-dummy.state
> Node: node-b (59295d90-5459-490d-a1e0-d48810cf2fb3): online
> Node: node-a (b3852a23-c10b-440a-a8e0-263b0185d657): online
>
> dummy   (ocf::heartbeat:Dummy): Started node-b
>
> Operations:
> * Node node-b:
>   dummy:
>    + start: rc=0 (ok)
>    + monitor: interval=10000ms rc=0 (ok)
> * Node node-a:
>   dummy:  fail-count=1
>    + start: rc=0 (ok)
>    + monitor: interval=10000ms rc=7 (not running)
>    + stop: rc=0 (ok)
>
> Failed actions:
>    dummy_monitor_10000 (node=node-a, call=4, rc=7): complete
>
>
>
> After that, the node which the resource is running now is stopped  
> manually.
>
> # service heartbeat stop
> Node: node-b (59295d90-5459-490d-a1e0-d48810cf2fb3): OFFLINE
> Node: node-a (b3852a23-c10b-440a-a8e0-263b0185d657): online
>
> Operations:
> * Node node-b:
>   dummy:
>    + start: rc=0 (ok)
>    + monitor: interval=10000ms rc=0 (ok)
>    + stop: rc=0 (ok)
> * Node node-a:
>   dummy:  fail-count=1
>    + start: rc=0 (ok)
>    + stop: rc=0 (ok)
>
> At this time, node-a's monitor NG disappears from crm_mon.

because it is no longer in the current start/stop series for the  
resource.

>
> It might be an expected behavior for now,

it is.

>
> it would be convenient if crm_mon can keep showing some past failures.

it cant display them forever.  they are not (and should not) be kept  
in the CIB forever as it would cause the CIB size to explode.





More information about the Pacemaker mailing list