Hi ,
this is not really true when grouping about with a clone, is it ?
and if we remove the colocation configuration, I think there will be no
failover of the group1 in the case clone-fs1 disappears , right ?
... but I've made some progress about this and there is something
strange that experts of Pacemaker configuration will undoublty explain
to me :
I have another set of resources collocated and ordered quite the same
way, and it
does the job as I want :
Clone Set: clone-fs-DBs
Started: [ node1 node2 ]
Resource Group: group-DBs
IPaddr-pgsqldb-server (ocf::heartbeat:IPaddr): Started node1
res-pgsql (ocf::heartbeat:pgsql): Started node1
you can see that the clone is started on both nodes, and this is quite
the same configuration as for group1 and clone-fs1, I have performed
quite the same crm configuration :
crm configure colocation coloc1-psql +INFINITY: group-DBs clone-fs-DBs
crm configure order order1-postgresql mandatory: clone-fs-DBs group-DBs
so this is quite the same configuration except that, to simplify my
previous ML
question I omitted to say that there were two clone-fs collocated and
ordered
with group1 : clone-fs1 and clone-fs2.
so in fact my colocation and order commands were :
crm configure colocation coloc1 +INFINITY: group1 clone-fs1 clone-fs2
crm configure order order1 mandatory: clone-fs1 clone-fs2 group1
so you see: quite the same command as for group-DBs but with two
clone-fs instead of one
and the behavior is not the same :
commands with only one clone-fs and a group, the clone-fs is well
started on both nodes
whatever node has started the group :
Clone Set: clone-fs-DBs
Started: [ node1 node2 ]
Resource Group: group-DBs
IPaddr-pgsqldb-server (ocf::heartbeat:IPaddr): Started node1
res-pgsql (ocf::heartbeat:pgsql): Started node1
commands with two clone-fs and a group, each clone-fs is started only on
the node where
the group is started
Clone Set: clone-fs1
Started: [ node1 ]
Stopped: [ fs-BCM-data:1 ]
Clone Set: clone-fs2
Started: [ node1 ]
Stopped: [ fs-BCM-log:1 ]
Resource Group: group1
IPaddr (ocf::heartbeat:IPaddr): Started node1
res-nagios (lsb:nagios): Started node1
So as commands are the same, and crm configure show does not reveal any
difference
between both configuration, I had a look in the cib.xml and the
difference is here :
with one clone-fs, the result of collocation and order crm config
commands is :
<rsc_colocation id="coloc1-psql" rsc="group-DBs" score="+INFINITY"
with-rsc="clone-fs-DBs"/>
<rsc_order first="clone-fs-DBs" id="order1-postgresql"
score="INFINITY" then="group-DBs"/>
whereas whith two clone-fs, the result of collocation and order crm
config commands is :
<rsc_colocation id="coloc1" score="+INFINITY">
<resource_set id="coloc1-0">
<resource_ref id="group1"/>
<resource_ref id="clone-fs1"/>
<resource_ref id="clone-fs2"/>
</resource_set>
</rsc_colocation>
<rsc_order id="order1" score="INFINITY">
<resource_set id="order1-0">
<resource_ref id="clone-fs1"/>
<resource_ref id="clone-fs2"/>
<resource_ref id="group1"/>
</resource_set>
</rsc_order>
In fact the behavior of the clone-fs-DBs and group-DBs seems more
logical (for me! ;-) ) as it is "clone" fs resources,
but the real question is : why the same colocation and/or order commands
leads to a different behavior,
depending if there is only 1 clone res or more than one in the
colocation and/or order crm commands ?
Thanks
Regards
Alain
> > crm configure colocation coloc1 +INFINITY: group1 clone-fs1
>
> This says that group1 and clone-fs1 have to be on the same machine. That
> prohibits starting clone-fs1 on a machine where group1 is not running.
> That isn't what you meant. I think all you need is the order directive
> to make sure clone-fs1 is started before group1.
>
> --Greg
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
