On 4/20/07, Bernd Schubert <[EMAIL PROTECTED]> wrote:
Hi,

when upgrading from heartbeat-2.0.5 to 2.0.8 another problem occured. We have
a script the admin can call to clean occured errors.

At the end of the script the date of the last cleaning action is set, using
the command "crm_attribute -n last_cleanup -v "`date -R`" ".
On heartbeat-2.0.5 this will change the attribute of the following
xml-fragment.

<cluster_property_set id="cib-bootstrap-options">
  <attributes>
    <nvpair id="cib-bootstrap-options-last_cleanup" name="last_cleanup"
value="Fri, 20 Apr 2007 15:39:57 +0200"/>
  </attributes>
</cluster_property_set>


Our initial xml file to set up heartbeat even doesn't contain the
id="cib-bootstrap-options-last_cleanup". So heartbeat-2.0.5 did create it
itself.
Unfortunately heartbeat-2.0.8 isn't doing that and so the crm_attribute
command fails.

the underlying bug that causes this command to fail has been fixed for
the next release.
i can dig up the patch if you need it


Unfortunately nobody remembers why it was set, but all are sure it was for a
special reason (**mental note for the future: always document this in the
scripts**).

its to prompt the cluster into running the PE again.

i'm guessing that "crm_resource -C" is mentioned a few lines earlier?

i'm pretty sure this is no longer required (as crm_resource will do
something equivalent automatically)



Is there a way to restore the old behaviour? I see in hb_gui there's also the
possibility clean up error states, how is hb_gui doing that?


Thanks in advance,
Bernd


--
Bernd Schubert
Q-Leap Networks GmbH
_______________________________________________
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