[Pacemaker] Collocating resource with a started clone instance

[Sergey Tachenov]

>>>> group postgres pgdrive_fs DBIP postgresql
>>>> colocation postgres_on_drbd inf: postgres ms_drbd_pgdrive:Master
>>>> order postgres_after_drbd inf: ms_drbd_pgdrive:promote postgres:start
>>>> ...
>>>> location DBIPcheck DBIP \
>>>>        rule $id="DBIPcheck-rule" 10000: defined pingd and pingd gt 0
>>>> location master-prefer-node1 DBIP 50: srvplan1
>>>> colocation DBIP-on-web 1000: DBIP tomcats
>>>
>>> try inf: ... 1000: will be not enough .... because DBIP is also part of
>>> postgres group and that group must follow the DRBD Master
>>
>> I understand, but why doesn't it move the DRBD Master (and the whole
>> group) then? If I set the score INF, then the DBIP gets stopped
>> completely, and ptest shows -INF for both nodes. And why the ping
>> location constraint works then? If the ping is not answered, the whole
>> group moves perfectly fine.
>
> hmm ... does it work better with a high value like 100000: ?

No. I tried various numbers, got this:

If I set the score to 1000, the DBIP score on srvplan1 is 9250.
If I set the score to 10000, the DBIP score on srvplan1 is 250.
If I set the score to 15000, the DBIP score on srvplan1 is... guess
what? 10250. The same for 100000.

The threshold seems to be 10150: with this value, the score is 100,
with 10251 it reverts to 10250. The score for srvplan2 remains at -INF
in all cases, though.