[Pacemaker] An odd behavior about a timeout for remote fencing operaions

[Satomi TANIGUCHI]

Hi,


I'm sending a patch which has the following roles.

(1) add a timeout in stonithd on DC node when it is going to broadcast
    a STONITH request because all STONITH plugins on DC failed.

(2) remove the timeout above when DC-stonithd receives a "STONITH succeeded"
    message from other node.

I have no idea whether the current behavior is intentional or a bug.
If it is an expected behavior, I'd like to hear its purpose.
If not, I hope this patch is helpful to you.

Best Regards,
Satomi TANIGUCHI


Satomi TANIGUCHI wrote:
> Hi Lists,
> 
> I found an odd behavior about remote fencing.
> 
> In the following case, stonithd on DC node adds the timeout for remote fencing
> operations.
> (1) the target node is DC.
> (2) the target node is not DC, but DC node has no STONITH plugin.
> 
> But it doesn't add that timeout in the following case.
> (i) the target node is not DC, and all STONITH plugins fail on DC node.
> 
> Is it an expected behavior or a bug?
> 
> When all STONITH operations on all nodes in the cluster failed,
> maybe no action can be executed forever because transition timeout in tengine
> is going to be removed...
> 
> Best Regards,
> Satomi TANIGUCHI
> 
> _______________________________________________
> Pacemaker mailing list
> Pacemaker at clusterlabs.org
> http://list.clusterlabs.org/mailman/listinfo/pacemaker

-------------- next part --------------
A non-text attachment was scrubbed...
Name: timeout_for_remote_fencing_operaions.patch
Type: text/x-patch
Size: 4454 bytes
Desc: not available
URL: <https://lists.clusterlabs.org/pipermail/pacemaker/attachments/20081114/5cb3b9d5/attachment-0001.bin>