On Mar 12, 2008, at 7:44 PM, Dejan Muhamedagic wrote:
Hi,
On Mon, Mar 10, 2008 at 10:52:19AM -0400, Doug Knight wrote:
All,
I am currently running heartbeat 2.0.8 (plans to upgrade are in the
works). I have a group of similar processes, currently configured as
ordered="true" and collocated="true", with the attributes within the
group tag in the XML. I have a group constraint to set the preferred
node for the group. I modified the ordered attribute of the group tag
from "true" to "false", leaving the collocated attribute set to
"true".
When I do that, the processes all start up asynchronously, as
expected,
but the processes are no longer collocated (in fact they evenly
distribute between the two systems, 5 one one, 4 on the other). Am I
mis-interpreting the ordered attribute?
No. It looks like a bug to me, but probably something which has
been fixed in the meantime.
Highly likely.
You should try to upgrade to 2.1.3.
At this point it would be more advisable to upgrade to pacemaker which
gets regular bugfix-only updates.
I saw in the online examples
that collocation there is done via an rsc_colocation constraint. Are
these two different ways of enforcing collocation?
Apparently, a group is just a shortcut for a set of order and
colocate constraints.
Nod. They're a syntactic shortcut to make upgrading from v1 easier.
Don't try and do too much with them.
The group definition
and constraint are shown below. If more information is needed
please let
me know.
<group ordered="false" collocated="true" id="grp_decoders">
...
<rsc_location id="locate_decoders" rsc="grp_decoders">
<rule id="rule_decoders_on_sysp" score="100">
<expression id="exp_decoders_on_sw" attribute="#uname"
operation="eq" value="sysp"/>
</rule>
</rsc_location>
This looks fine.
Thanks,
Dejan
Thanks,
Doug Knight
WSI Corp
Andover, MA
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems