On 7/18/07, Taldevkar, Chetan <[EMAIL PROTECTED]> wrote:

Hi,

It seems there is no way I can give equal score to nodes and allow
resource to run on one of the node based on the return code of the start
operation.

Of course not.


What I found is Linux-HA tries to start resource on both the nodes if
start operation fails on one of the node then it does not fail over on
that node again.

right - by design.
otherwise "bad" resources just end up ping-ponging around the cluster

If I try to use crm_resource -C to clean the operation on failed node it
cleans up and immediately tries to execute start operation on the failed
node.

right - thats what -C is for... clearing away old (and no longer
relevant errors)

Is there any way we can stop immediate start of resource on failed node
after cleanup of failed node through crm_resource.

set the resource's target_role to stopped
crm_resource -p target_role --meta -v stopped -r your_rsc_id
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to