On 7/18/07, Maxim Veksler <[EMAIL PROTECTED]> wrote:
On 7/18/07, Adrian Chapela <[EMAIL PROTECTED]> wrote:
> Hello,
>
> I am reading list and documentation but I don't understand
> resource_stickiness, default_resource_stickiness ,
> default_resource_failure_stickiness.
>
I myself would appreciate a more experienced answer from the gurus of the list.
For me, I have 5 resources and I want stickiness the same way you do -
If resource fails (even 1 time) on a node, I want all my resource
group to fail over to another node suitable for running this group. I
also don't want my resource groups to fail back to the original node
if it becomes "suitable" once again (to avoid unnecessary ping-pong).
The only way I could achieve this is by setting :
<nvpair id="default-resource-stickiness"
name="default-resource-stickiness" value="21"/>
<nvpair id="default-resource-failure-stickiness"
name="default-resource-failure-stickiness" value="-100"/>
The calculation is :
NODE-Stickiness = 5 * 21 (5 resources in my group), i.e
if (105 / abs(100)) > 1; then...
I'm seeing this works this way on heartbeat 2.0.8, with failcount of 2
and sometimes even 3.
I couldn't find a clear answer as to why it works this way and how I
can have exactly 1 failure before failover.
if you want it to move immediately, then you need:
default-resource-failure-stickiness > NODE-Stickiness
so setting default-resource-failure-stickiness = 106 (should) do what you want.
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
