On 16/05/2013, at 7:57 PM, Guglielmo Abbruzzese <[email protected]> wrote:
> Hi,
> I'd like to hear from you if someone has already experienced someting
> similar or -in case- to get how to do better.
> I need to sync a few files from time to time. Non need for a storage or a
> DRBD solution. Active/Passive cluster,
>
> I created a OCF master/slave resource from the template called "syncer",
> adding in the RA_syncer::monitor() method a couple of rsync commands as the
> following example:
> ...
> "/usr/bin/rsync -avz --delete ${SOURCE_CFG_FOLDER_X}/
> ${OTHER_NODE}::ALIAS_X"
> ...
> The ocf-tester says the RA is ok.
>
> Then I add a constraint to pacecemaker, so that the resource_ms_syncer is
> Master only if the other resources are master on the same node.
>
> I have performed several tests, and the sync seems to work fine either on
> node A or on node B if I perform some swaps in order to validate the
> solution.
>
> But in one case something goes wrong: while A is master, if I shut down the
> B node after a while the RA::monitor timer on the A node seems to stop
> working. Of course I expect to get some OCF log errors such as
> "(resource_syncer:0:monitor:stderr) rsync: failed to connect to NODE_B", but
> when I restart the NODE_B I would like the resync to start working again. To
> make it so I must restart the HA services on the NODE_A, and this is not
> acceptable :(
Modify the monitor operation of your RA to test if the other node is available
again and effectively perform a start action?
>
> On the footer of this email I have attached the cib section for the ocf
> resource I created. Could the problem be related to some timeout properties
> I failed to set? Any suggestion?
>
> Thanks a lot
> G.
>
> ----
>
> COMPONENTS:
> OS: RHEL6.2 2.6.32-220.el6.x86_64
> Pacemaker: pacemaker-1.1.6-3.el6.x86_64 (in bundle with the OS)
> Corosync: corosync-1.4.1-4.el6.x86_64 (in bundle with the OS)
> Rsync: rsync-3.0.6-5.el6_0.1.x86_64
>
>
> RESOURCE:
> <master id="resource_ms_syncer">
> <meta_attributes id="resource_ms_syncer-meta_attributes">
> <nvpair id="resource_ms_syncer-meta_attributes-master-max"
> name="master-max" value="1"/>
> <nvpair id="resource_ms_syncer-meta_attributes-master-node-max"
> name="master-node-max" value="1"/>
> <nvpair id="resource_ms_syncer-meta_attributes-clone-max"
> name="clone-max" value="2"/>
> <nvpair id="resource_ms_syncer-meta_attributes-clone-node-max"
> name="clone-node-max" value="1"/>
> <nvpair id="resource_ms_syncer-meta_attributes-notify"
> name="notify" value="true"/>
> <nvpair id="resource_ms_syncer-meta_attributes-target-role"
> name="target-role" value="Started"/>
> </meta_attributes>
> <primitive class="ocf" id="resource_syncer" provider="resi"
> type="syncer">
> <instance_attributes id="resource_syncer-instance_attributes">
> <nvpair id="resource_syncer-instance_attributes-state"
> name="state" value="/var/run/resource_syncer.state"/>
> <nvpair
> id="resource_syncer-instance_attributes-internal_parameter"
> name="internal_parameter" value="idle"/>
> </instance_attributes>
> <operations>
> <op id="resource_syncer-startup_M" interval="30s" name="monitor"
> role="Master"/>
> <op enabled="false" id="resource_syncer-startup_S"
> interval="40s" name="monitor" on-fail="restart" requires="nothing"
> role="Slave" timeout="60s"/>
> <op id="resource_syncer-start-0" interval="0" name="start"
> timeout="80s"/>
> <op id="resource_syncer-stop-0" interval="0" name="stop"
> timeout="80s"/>
> </operations>
> </primitive>
>
> _______________________________________________
> 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
