[ClusterLabs] [OCF] Pacemaker reports a multi-state clone resource instance as running while it is not in fact

Keisuke MORI keisuke.mori+ha at gmail.com
Thu Jan 7 00:17:10 EST 2016


2016-01-06 22:57 GMT+09:00 Jan Pokorný <jpokorny at redhat.com>:
> Hello ,
> On 04/01/16 17:33 +0100, Bogdan Dobrelya wrote:

>> Note, that it seems the very import action causes the issue, not the
>> ocf_run or ocf_log code itself.
>> [0] https://github.com/ClusterLabs/resource-agents/issues/734
> Have to wonder if there is any correlation with the issue discussed
> recently:
> http://oss.clusterlabs.org/pipermail/users/2015-November/001806.html
> Note that ocf:pacemaker:ClusterMon resource also sources
> ocf-shellfuncs set of helper functions.

I think that it's unlikely to be relevant.

In the case of this time, /bin/dash is apparently the cause of the
fork bomb by a bug in handling PS4 shell variable from the result of
testing on the github issue 734.

In the case of ClusterMon you mentioned, crm_mon is the one forking
repeatedly, not the ClusterMon shell script, not invoked via /bin/dash

Keisuke MORI

More information about the Users mailing list