Binoy,
You can disable application-level recovery actions by suspending all
resource groups:
clrs suspend +
The resource groups and resources would remain in their current state,
but no automatic RG failovers or restarts would occur even if there are
probe failures. You can still manually switch resources or groups using
commands such as "clresource disable", "clresource enable",
"clresourcegroup switch", clresourcegroup offline", etc.
However, this command suspends cluster activity only at the application
level. The lower level cluster services (heartbeats, device services,
...) continue to run. As Detlef says, we don't yet have a way to take a
node out of the cluster without evacuating services from it and without
rebooting.
On 02/23/10 12:17 AM, Detlef Ulherr wrote:
> Hi Binoy,
>
> all you can do is to evacuate the node with clnode vacuate <nodename>
> and then you can do whatever you want with the evacuated node. In
> example rebootin out of cluster mode. With clnode evacuate you have
> migrated all device groups and applications from this node.
>
> The cluster itself remains running on this node after evacuation, but
> it is ideling.
>
> The only way to stop the complete cluster service on a node is to
> evacuate it and reboot out of cluster mode.
>
> HTH Detlef
>
>
> binoy wrote:
>> Hi All,
>>
>> please give me some clarificarion about my concerns on sun cluster,
>>
>> Is it possible to stop sun cluster service on a particular sun
>> cluster node ?
>> When i tried the command scshutdown,all cluster nodes moved to
>> shutdown state.So how can i stop the sun cluster service on a
>> particular node without effecting the node state(i mean,with out
>> reebooting).
>>
>> All answers are highly appreciated.
>>
>> Regards,
>> Binoy
>>
>
