[Pacemaker] About replacement of clone and handling of the fail number of times.

[Andrew Beekhof]

On Fri, Mar 26, 2010 at 2:54 AM,  <renayama19661014 at ybb.ne.jp> wrote:
> Hi Andrew,
>
>> globally-unique=false" means that :0 and :1 are actually the same resource.
>> its perfectly valid for entries for both to exist on the node, but the
>> PE should fold them together internally.
>>
>> in most ways it does, just not for failures (yet).
>
> Thank you for comment.
> Some we were confused.
>
> In the first place, by setting of "globally-unique", what kind of difference is there?

Its a cue for the PE to turn on some optimizations for the clone.
For example, there's no point asking each node if it has instances 1-N
active since only 1 can ever be running.

This makes writing the monitor function a little easier too.
Check what ocf:pacemaker:ping does based on the globally-unique setting.

> In addition, what kind of place do you use it in?

I use it for any clone that can only be active once per machine, ie.
named, drbd, apache.
Ie. if the monitor function cannot tell the difference between two
instances of a clone, then globally-unique is a must.


Btw. Can you update your address book to use the correct list address please?
You still seem to be using @clusterlabs.org (instead of @oss.clusterlabs.org).