Hi,

On Thu, Jul 08, 2010 at 02:25:02PM +0200, Trujillo Carmona, Antonio wrote:
> 
> I'm testing a HA failover and right now all thing look fine but I found
> same fail, if I unplug the wire it change the node, when I plug it it
> return to the default, that is ok.
> If I stop the other service (haproxy) it restart it, but if I kill the
> service in a way it cannot be restart it put the service in "unmanaged
> in that node but don't change the node.
> this is my configuration and status on fail:
> =============================================================================
> balanceador-1:~# crm
> crm(live)# status
> ============
> Last updated: Thu Jul  8 14:22:29 2010
> Stack: openais
> Current DC: balanceador-1 - partition with quorum
> Version: 1.0.8-f2ca9dd92b1d+ sid tip
> 2 Nodes configured, 2 expected votes
> 2 Resources configured.
> ============
> 
> Online: [ balanceador-1 balanceador-2 ]
> 
>  Resource Group: weblogic
>      failover-ip      (ocf::heartbeat:IPaddr2):       Started balanceador-1
>      failover-haproxy (lsb:haproxy):  Started balanceador-1 (unmanaged)
> FAILED
>  Clone Set: pingdclone
>      Started: [ balanceador-1 balanceador-2 ]
> 
> Failed actions:
>     failover-haproxy_monitor_15000 (node=balanceador-1, call=197, rc=7,
> status=complete): not running
>     failover-haproxy_stop_0 (node=balanceador-1, call=198, rc=4,
> status=complete): insufficient privileges

You have to fix this. There's probably more info in the logs. No
failover is going to happen unless a resource is stopped. You're
also missing stonith: if you had one, this node would be fenced
and failover would then proceed.

> crm(live)# 
> =============================================================================
> crm(live)configure# show
> node balanceador-1 \
>       description="nodo fisico" \
>       attributes standby="false"
> node balanceador-2 \
>       description="nodo virtual" \
>       attributes standby="false"
> primitive failover-haproxy lsb:haproxy \
>       op monitor interval="15s"
> primitive failover-ip ocf:heartbeat:IPaddr2 \
>       params ip="10.104.16.234" lvs_support="true"
> unique_clone_address="true" \
>       op monitor interval="10s" enabled="true" on-fail="restart"
> primitive pingd ocf:pacemaker:pingd \
>       params host_list="10.104.24.204" multiplier="100" \
>       op monitor interval="15s" timeout="5s"
> group weblogic failover-ip failover-haproxy \
>       meta is-managed="true" target-role="started"
> clone pingdclone pingd \
>       meta globally-unique="false"
> location cli-prefer-weblogic weblogic \
>       rule $id="cli-prefer-rule-weblogic" $role="Master" 0: #uname eq
> balanceador-1
> xml <rsc_location id="weblogic_en_nodo_con_red" rsc="weblogic"> \
>       <rule id="weblogic_en_nodo_con_red-rule" score="-INFINITY"> \
>               <expression attribute="pingd"
> id="weblogic_en_nodo_con_red-expression-0" operation="lte" value="0"/> \
>               <expression attribute="pingd" 
> id="weblogic_en_nodo_con_red-rule-pingd"
> operation="not_defined" value="or"/> \

You have an error here. I guess that you wanted to say:

rule $id="weblogic_en_nodo_con_red-rule" -inf: pingd lte 0 or not_defined pingd

The easiest way to fix it is to first edit the constraint and
remove the string 'value="or"', so that crm can show it in its
notation. Then edit the constraint again and fix the expression.

Thanks,

Dejan

>       </rule> \
> </rsc_location>
> property $id="cib-bootstrap-options" \
>       dc-version="1.0.8-f2ca9dd92b1d+ sid tip" \
>       cluster-infrastructure="openais" \
>       expected-quorum-votes="2" \
>       last-lrm-refresh="1278590317" \
>       stonith-enabled="false" \
>       node-health-strategy="migrate-on-red" \
>       no-quorum-policy="ignore"
> op_defaults $id="op_defaults-options" \
>       record-pending="true"
> =============================================================================
> usualy I use pacemaker_gui
> Thank
> -- 
> Por favor, NO utilice formatos  de archivo  propietarios para el
> intercambio de  documentos, como DOC y XLS, sino HTML, RTF, TXT,
> CSV o cualquier otro que no obligue a utilizar un programa de un
> fabricante  concreto para tratar la información contenida en él.
> SALUD.
> _______________________________________________
> 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