Hello,
I am having some problems trying to configure a cluster made of 3 servers.
Server 1 - resources sql http sip +ip0
Server 2 - resources sql http sip +ip1
Server 3 - resources sip +ip3
What I did since sip depends on sql being started and also the ipX depends on
sip started is created a group between:
http+sql next created a clone of this group (max-clones=2 and only one per
server) with location restrictions on server 1 +2 and -infinity on 3 (let's
call this clonesql)
Created a clone with sip (let's call this clonesip) with max-clones=3 and only
1 per server
Also created a collocation constraint for ip0-3 to start only on clonesip.
I've put the whole config bellow.
Tested scenarios:
Changed on server 3 init script to return 1 on start and stopped sip. The ip3
moved to server 2. And this was correct.
Then started it again.
Next stopped sql on server 2, since this caused sip on server 3 to stopped (sip
on server 3 was connected to sql on server2) this caused an undesired
behavior. Instead of moving all the ips on server on it stopped them all,
Also clone-sip stopped working on all servers: (I might have done this two
times to end up here)
My problem is what am i doing wrong here, since if i understood the manual
correctly then this shouldn't be happening. Also if I try similar configuration
with collocation_set or order_set it does not take them into account.
I'm using the public rpms from opensuse repository pacemaker-1.0.3-2.2 for
centos 5 x86_64
Here is crm_mon output:
Last updated: Tue May 26 14:55:52 2009
Current DC: registration1 - partition with quorum
3 Nodes configured, 2 expected votes
5 Resources configured.
============
Online: [ registration2 registration3 registration1 ]
Clone Set: clonesql
Resource Group: voip:2
failover-sql:2 (lsb:mysql): ORPHANED Stopped
failover-http:2 (lsb:voip): ORPHANED Started registration1
Started: [ registration1 ]
Stopped: [ voipnow:0 ]
Failed actions:
failover-sql:0_monitor_30000 (node=registration2, call=55, rc=7,
status=complete): not running
failover-sip:2_monitor_15000 (node=registration2, call=76, rc=7,
status=complete): not running
failover-sip:0_monitor_15000 (node=registration2, call=86, rc=7,
status=complete): not running
failover-ip0_start_0 (node=registration3, call=13, rc=1, status=complete):
unknown error
failover-sip:1_monitor_15000 (node=registration3, call=1717, rc=7,
status=complete): not running
#default vars:
crm_attribute --attr-name stonith-enabled --attr-value false
crm_attribute --attr-name stop-orphan-resources --attr-value true
crm_attribute --attr-name stop-orphan-actions --attr-value true
<constraints>
<rsc_order id="order-sip" first="clonesip" score="INFINITY" then="clonesql"/>
<rsc_location id="loc-1" rsc="clonesql" node="registration1" score="200"/>
<rsc_location id="loc-2" rsc="clonesql" node="registration2" score="200"/>
<rsc_location id="loc-3" rsc="clonesql" node="registration3" score="-INFINITY"/>
<rsc_location id="loc-4" rsc="failover-ip0" node="registration1" score="15000"/>
<rsc_location id="loc-5" rsc="failover-ip0" node="registration2" score="1000"/>
<rsc_location id="loc-6" rsc="failover-ip0" node="registration3"
score="-INFINITY"/>
<rsc_location id="loc-7" rsc="failover-ip1" node="registration1" score="1000"/>
<rsc_location id="loc-8" rsc="failover-ip1" node="registration2" score="15000"/>
<rsc_location id="loc-9" rsc="failover-ip1" node="registration3"
score="-INFINITY"/>
<rsc_location id="loc-10" rsc="failover-ip2" node="registration1" score="100"/>
<rsc_location id="loc-11" rsc="failover-ip2" node="registration2" score="1000"/>
<rsc_location id="loc-12" rsc="failover-ip2" node="registration3"
score="15000"/>
<rsc_colocation id="ip0-colo" rsc="failover-ip0" score="INFINITY"
with-rsc="clonesip"/>
<rsc_colocation id="ip1-colo" rsc="failover-ip1" score="INFINITY"
with-rsc="clonesip"/>
<rsc_colocation id="ip2-colo" rsc="failover-ip2" score="INFINITY"
with-rsc="clonesip"/>
</constraints>
<resources>
<primitive class="ocf" id="failover-ip0" provider="heartbeat"
type="IPaddr">
<instance_attributes id="failover-ip0-instance_attributes">
<nvpair id="failover-ip0-instance_attributes-ip" name="ip"
value="192.168.1.3"/>
</instance_attributes>
<operations>
<op id="failover-ip0-monitor-10s" interval="10s" name="monitor"
on-fail="stop"/>
</operations>
<meta_attributes id="failover-ip0-meta_attributes">
<nvpair id="failover-ip0-meta_attributes-target-role"
name="target-role" value="Started"/>
</meta_attributes>
</primitive>
<primitive class="ocf" id="failover-ip1" provider="heartbeat"
type="IPaddr">
<instance_attributes id="failover-ip1-instance_attributes">
<nvpair id="failover-ip1-instance_attributes-ip" name="ip"
value="192.168.1.5"/>
</instance_attributes>
<operations>
<op id="failover-ip1-monitor-10s" interval="10s" name="monitor"
on-fail="stop"/>
</operations>
<meta_attributes id="failover-ip1-meta_attributes">
<nvpair id="failover-ip1-meta_attributes-target-role"
name="target-role" value="Started"/>
</meta_attributes>
</primitive>
<primitive class="ocf" id="failover-ip2" provider="heartbeat"
type="IPaddr">
<instance_attributes id="failover-ip2-instance_attributes">
<nvpair id="failover-ip2-instance_attributes-ip" name="ip"
value="192.168.1.7"/>
</instance_attributes>
<operations>
<op id="failover-ip2-monitor-10s" interval="10s" name="monitor"
on-fail="stop"/>
</operations>
<meta_attributes id="failover-ip2-meta_attributes">
<nvpair id="failover-ip2-meta_attributes-target-role"
name="target-role" value="Started"/>
</meta_attributes>
</primitive>
<clone id="clonesip">
<meta_attributes id="clonesip-meta_attributes">
<nvpair id="clonesip-meta_attributes-clone-max" name="clone-max"
value="3"/>
<nvpair id="clonesip-meta_attributes-globally-unique"
name="globally-unique" value="false"/>
<nvpair id="clonesip-meta_attributes-clone-node-max"
name="clone-node-max" value="1" />
</meta_attributes>
<primitive class="lsb" id="failover-sip" type="kamailio">
<operations>
<op id="failover-sip-monitor-15s" interval="15s" name="monitor"
on-fail="stop"/>
</operations>
<meta_attributes id="failover-sip-meta_attributes">
<nvpair id="failover-sip-meta_attributes-target-role"
name="target-role" value="Started"/>
</meta_attributes>
</primitive>
</clone>
<clone id="clonesql">
<meta_attributes id="clonesql-meta_attributes">
<nvpair id="clonesql-meta_attributes-clone-max" name="clone-max"
value="2"/>
<nvpair id="clonesql-meta_attributes-globally-unique"
name="globally-unique" value="false"/>
<nvpair id="clonesql-meta_attributes-clone-node-max"
name="clone-node-max" value="1" />
</meta_attributes>
<group id="voip" >
<meta_attributes id="voip-meta_attributes">
<nvpair id="voip-meta_attributes-collocated" name="collocated"
value="true"/>
</meta_attributes>
<primitive class="lsb" id="failover-sql" type="mysql">
<operations>
<op id="failover-sql-start" name="start" interval="0" timeout="30s"/>
<op id="failover-sql-stop" name="stop" interval="0" timeout="30s"/>
<op id="failover-sql-monitor-30s" interval="30s" name="monitor"
on-fail="stop"/>
</operations>
<meta_attributes id="failover-sql-meta_attributes">
<nvpair id="failover-sql-meta_attributes-target-role"
name="target-role" value="Started"/>
</meta_attributes>
</primitive>
<primitive class="lsb" id="failover-http" type="voip">
<operations>
<op id="failover-http-monitor-15s" interval="15s" name="monitor"
on-fail="stop"/>
</operations>
<meta_attributes id="failover-http-meta_attributes">
<nvpair id="failover-http-meta_attributes-target-role"
name="target-role" value="Started"/>
</meta_attributes>
</primitive>
</group>
</clone>
</resources>
_______________________________________________
Pacemaker mailing list
Pacemaker@oss.clusterlabs.org
http://oss.clusterlabs.org/mailman/listinfo/pacemaker
