On Fri, March 20, 2009 23:33, adam wrote:
> Andreas-
>
>
>
> Actually, the opposite.  Sorry if first email was unclear.  Dummy_A goes
> to juliet after killing heartbeat on kilo, but fails back, regardless of
> score, when kilo returns.  After going down twice, Dummy_A should prefer
> juliet over kilo.

ahhh ... ok ... after rethinking the situation I would guess the
state-file of Dummy_A is still there on kilo in
"/var/run/heartbeat/rsctmp/" and on a restart of heartbeat this is
interpreted as multiple-active. The default policy for this is to stop all
and restart one and then the the higher score of kilo wins ... have a look
at your logs

... btw resource_failure_stickiness is only effective for
resource-failures and not node-failures but as Andrew already mentioned:
use Pacemaker and its "migration-treshhold"

Regards,
Andreas

>
> auto_failback is set to off in ha.cf, too.

ignored when using crm

>
> Thanks again
>
>
> You mean Dummy_A stays on juliet after forcibly killing and restarting
> heartbeat on kilo?
>
>
> Expected because 150+100 > 200 ... you disabled auto-failback with the
> resource_stickiness.
>
> You can use 'ptest' to examine the scores.
>
>
> Regards,
> Andreas
>
>
>>
>> Thanks in advance.
>>
>>
>>
>> <resources>
>> <primitive class="ocf" id="Dummy_A"
>>
> provider="heartbeat" type="Dummy">
>>     <meta_attributes id="dummy-ma">
>>       <attributes>
>>            <nvpair name="target_role"
>>
> id="dummystarted" value="started"/>
>>            <nvpair name="resource_stickiness"
>>
> id="ma-1" value="100"/>
>>            <nvpair name="resource_failure_stickiness"
>>
> id="ma-2"
>> value="-100"/>       </attributes>     </meta_attributes>
>> </primitive>
>> </resources>
>> <constraints>
>> <rsc_location id="dummy_on_kilo" rsc="Dummy_A">
>>     <rule id="kilo_native" score="200">
>>         <expression id="be_on_kilo"
>>
> attribute="#uname" operation="eq"
>> value="kilo"/>     </rule>        <rule id="juliet_backup"
>>
> score="150">
>>                 <expression
>>
> id="fail_to_juliet" attribute="#uname"
>> operation="eq" value="juliet"/>       
> </rule>
>
>> </rsc_location>
>> </constraints>
>>
>>
>>
>>
>>
>>
>>
>> _______________________________________________
>> Linux-HA mailing list
>> [email protected]
>> http://lists.linux-ha.org/mailman/listinfo/linux-ha
>> See also: http://linux-ha.org/ReportingProblems
>>
>>
>>
>>
>
>
> --
> : Andreas Kurz
> : LINBIT | Your Way to High Availability
> : Tel +43-1-8178292-64, Fax +43-1-8178292-82
> :
> : http://www.linbit.com
>
>
> DRBD® and LINBIT® are registered trademarks of LINBIT, Austria.
>
>
> This e-mail is solely for use by the intended recipient(s). Information
> contained in this e-mail and its attachments may be confidential,
> privileged or copyrighted. If you are not the intended recipient you are
> hereby formally notified that any use, copying, disclosure or distribution
> of the contents of this e-mail, in whole or in part, is prohibited. Also
> please notify immediately the sender by return e-mail and delete this
> e-mail from your system. Thank you for your co-operation.
>
> _______________________________________________
> 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

Reply via email to