Jaspal
I think due to terminology (or naming ) of this node makes head spin.  
if you put nodeA, NodeB and nodeC for explanation it will make more sense.

I had similar type of problem and I think it is due to ha version. ( we are 
also using 2.1.3 )
I have four node setup when two node dies whole setup gets confuse. I think 
something to do with quorum.
you can use some kind of constrain to see if it works.
Also try to use resource group. 


> Date: Tue, 5 Jan 2010 22:58:27 +0530
> From: [email protected]
> To: [email protected]
> Subject: Re: [Linux-HA] 3 node clustering issue while both of Active node     
> fail their resources same time on to passive
> 
> Hello,
> 
> Hi,
> 
> On Tue, Jan 05, 2010 at 07:54:30PM +0530, jaspal singla wrote:
> > Hello,
> >
> > I have 3 nodes cluster with 2 active and 1 passive nodes.
> >
> > I have configured 2 resources groups containing all necessary file system,
> > ip address and vz-script.
> >
> > My cluster works perfect if both of Active nodes fail there resources one
> at
> > a time to passive node but the issue which I am facing is when I try to
> fail
> > both of the active node at the same time one by one then I noticed that
> the
> > 2nd active node which fails there resources after the 1st active node stop
> > the resources of 1st active node  from passive node(which presently
> running
> > on passive node) and start its own resources over the passive node which
> > should not be there.
> 
> Could you please try with shorter sentences, I got lost half way. Apart from
> that, please upgrade from 2.1.3 to pacemaker.  2.1.3 is old and you won't
> get much support for it.
> 
> Actually I am already spent lot of time with 2.1.3 and  Pacemaker is totally
> new 4 me and I have done load of R & D on 3 node clustering with 2.1.3
> version. And also my management wants me to live this setup as soon as
> possible. So I am just wondering for any help on 2.1.3 version.
> 
> My 3 node cluster is working fine if I test the cluster by getting down my
> active nodes one by one.
> 
> But this issue come when my both Active machine gets down simultaneously one
> after another.
> 
> Problem is:
> 
> In case of node_master active node fails and failover its resources to
> node_slave passive node (Problematic scenario):-
> If node_master fail there resources on node_slave then there isn't any
> problem for that but at the same time if node3 fail there resources then
> node_master resources stops on node_slave and node3 resources starts over
> node_slave node.
> 
> But whereas,
> 
> In case of node3 active node fails and failover its resources to node_slave
> passive node (No problem with this scenario):-
> If node3 fail there resources on node_slave then there isn't any problem and
> at the same time if node_master fail there resources on node_slave then
> node3 finds that node_slave already have occupied with resources and stop
> node3 resources. (Perfect scenario)
> 
> Any help will be greatly appreciated!
> 
> Thanks,
> 
> Jaspal Singla
> 
> > As I think the 2nd active node should stop there resources because passive
> > node is already occupied with the 1st active node resources.
> >
> >
> > Is that normal? And how I avoid not to stop the resources which already
> run
> > on passive node by the last active node which fails.
> >
> >
> > I am just explaining my scenario below, hope this will give better idea to
> > understand the exact issue is and also pasting my cib.xml file of my
> cluster
> > below.
> >
> > The 3 mode are as:
> >
> > 1) node_master ---> Active node
> > 2) node_slave  ---> Passive node
> > 3) node3 ---------> Active node
> >
> > Problem:
> >
> > If node_master fail there resources on node_slave then there isn't any
> > problem for that but at the same time if node3 fail there resources then
> > node_master resources stops on node_slave and node3 resources starts over
> > node_slave node. (Problematic scenario)
> >
> > But whereas,
> >
> > If node3 fail there resources on node_slave then there isn't any problem
> and
> > at the same time if node_master fail there resources on node_slave then
> > node3 fi
> > nds that node_slave already have occupied with resources and stop node3
> > resources. (Perfect scenario)
> > Any help will be highly appreciated.
> >
> > The details for the my scenerio are as:
> >
> > 1) I am running CentOS-5.3
> > 2) Kernel version is 2.6.18-028stab059.6 (Virtoozzo Kernel)
> > 3) Heartbeat version- 2.1.3-3.
> > 4) Installation through RPM.
> > 5) Using Heartbeat version 2.
> > 6) cat /etc/ha.d/ha.cf
> > ______________________________
> >
> > _____________________________________________________________
> > >
> > > deadtime  10
> > > bcast eth1
> > > crm yes
> > > node node_master
> > > node node_slave
> > > node node3
> > > debugfile /var/log/ha-debug
> > > logfile /var/log/ha-log
> > > logfacility     local0
> > >
> > >
> > _____________________________________________________________________________________________
> > >
> > > This file is same in all the 3 nodes.
> > >
> > > 7) cib.xml
> > >
> > >
> > -------------------------------------------------------------------------------------------------------
> > > <cib admin_epoch="0" have_quorum="true" ignore_dtd="false" num_peers="3"
> > > cib_feature_revision="2.0" crm_feature_set="2.0" ccm_transition="3"
> > > generated="true" dc_uuid="d90b1ed2-0000-44ac-9a4d-b435a6befd36"
> > epoch="291"
> > > num_updates="2" cib-last-written="Mon Jan  4 19:52:33 2010">
> > >    <configuration>
> > >      <crm_config>
> > >        <cluster_property_set id="cib-bootstrap-options">
> > >          <attributes>
> > >            <nvpair id="cib-bootstrap-options-dc-version"
> > name="dc-version"
> > > value="2.1.3-node: 552305612591183b1628baa5bc6e903e0f1e26a3"/>
> > >            <nvpair id="cib-bootstrap-options-symmetric-cluster"
> > > name="symmetric-cluster" value="false"/>
> > >            <nvpair name="last-lrm-refresh"
> > > id="cib-bootstrap-options-last-lrm-refresh" value="1262555178"/>
> > >          </attributes>
> > >        </cluster_property_set>
> > >      </crm_config>
> > >      <nodes>
> > > <node uname="node3" type="normal"
> > id="7e5fdac9-80dc-41a7-bd8f-a5591a1b69a0">
> > >          <instance_attributes
> > > id="nodes-7e5fdac9-80dc-41a7-bd8f-a5591a1b69a0">
> > >            <attributes>
> > >              <nvpair name="standby"
> > > id="standby-7e5fdac9-80dc-41a7-bd8f-a5591a1b69a0" value="on"/>
> > >            </attributes>
> > >          </instance_attributes>
> > >        </node>
> > >        <node uname="node_slave" type="normal"
> > > id="d90b1ed2-0000-44ac-9a4d-b435a6befd36">
> > >          <instance_attributes
> > > id="nodes-d90b1ed2-0000-44ac-9a4d-b435a6befd36">
> > >            <attributes>
> > >              <nvpair name="standby"
> > > id="standby-d90b1ed2-0000-44ac-9a4d-b435a6befd36" value="off"/>
> > >            </attributes>
> > >          </instance_attributes>
> > >        </node>
> > >        <node uname="node_master" type="normal"
> > > id="075961d1-4492-4ba9-b4ad-e8c27b9e3f4b">
> > >          <instance_attributes
> > > id="nodes-075961d1-4492-4ba9-b4ad-e8c27b9e3f4b">
> > >            <attributes>
> > >              <nvpair name="standby"
> > > id="standby-075961d1-4492-4ba9-b4ad-e8c27b9e3f4b" value="on"/>
> > >            </attributes>
> > >          </instance_attributes>
> > >        </node>
> > >      </nodes>
> > >      <resources>
> > >        <group id="group_vz_1">
> > >          <meta_attributes id="group_vz_1_meta_attrs">
> > >  <attributes>
> > >              <nvpair name="target_role"
> > id="group_vz_1_metaattr_target_role"
> > > value="started"/>
> > >              <nvpair id="group_vz_1_metaattr_ordered" name="ordered"
> > > value="true"/>
> > >              <nvpair id="group_vz_1_metaattr_collocated"
> > name="collocated"
> > > value="true"/>
> > >              <nvpair id="group_vz_1_metaattr_resource_stickiness"
> > > name="resource_stickiness" value="900"/>
> > >            </attributes>
> > >          </meta_attributes>
> > >          <primitive id="resource_ipaddr" class="ocf" type="IPaddr"
> > > provider="heartbeat">
> > >            <instance_attributes id="resource_ipaddr_instance_attrs">
> > >              <attributes>
> > >                <nvpair id="69fd1897-4ec1-4bdb-a33f-1d2bf1eda0da"
> > name="ip"
> > > value="66.199.245.207"/>
> > >  <nvpair id="177a2461-1b7a-4875-a866-916ed730b396" name="nic"
> > value="eth0"/>
> > >                <nvpair id="347bf250-d25f-41df-bb7f-36f273c8029c"
> > > name="cidr_netmask" value="255.255.255.224"/>
> > >              </attributes>
> > >            </instance_attributes>
> > >            <operations/>
> > >          </primitive>
> > >          <primitive id="resource_filesystem" class="ocf"
> > type="Filesystem"
> > > provider="heartbeat">
> > >            <instance_attributes id="resource_filesystem_instance_attrs">
> > >              <attributes>
> > >                <nvpair id="11aaf508-8f1f-4a9d-adde-2ff7e6a82740"
> > > name="device"
> > > value="/dev/disk/by-uuid/f5feb406-685a-41f8-a4f7-170ae0925901"/>
> > >                <nvpair id="517a54cf-70cd-46bf-8ae6-0953d3617599"
> > > name="directory" value="/vz"/>
> > >                <nvpair id="b0a7cb9d-d0be-45fe-afcb-2860745bc5d5"
> > > name="fstype" value="ext3"/>
> > >                <nvpair id="06344316-a2c5-4ced-930f-5e151dfbe1e2"
> > > name="options" value="_netdev,noatime"/>
> > >              </attributes>
> > >            </instance_attributes>
> > >            <operations/>
> > >          </primitive>
> > >          <primitive id="resource_vz_script_1" class="lsb" type="vz"
> > > provider="heartbeat">
> > >            <operations>
> > >              <op id="88f60be0-19ac-4dd1-bbf3-0471a4c7bd03" name="monitor"
> > > interval="15s" timeout="30s" start_delay="0s" on_fail="restart"/>
> > >              <op id="387958aa-3ede-46c0-b77c-495e6cd44192" name="stop"
> > > timeout="200s"/>
> > >            </operations>
> > >          </primitive>
> > >        </group>
> > >        <group id="group_vz_2">
> > >  <meta_attributes id="group_vz_2_meta_attrs">
> > >            <attributes>
> > >              <nvpair name="target_role"
> > id="group_vz_2_metaattr_target_role"
> > > value="started"/>
> > >              <nvpair id="group_vz_2_metaattr_ordered" name="ordered"
> > > value="true"/>
> > >              <nvpair id="group_vz_2_metaattr_collocated"
> > name="collocated"
> > > value="true"/>
> > >              <nvpair name="resource_stickiness"
> > > id="group_vz_2_metaattr_resource_stickiness" value="900"/>
> > >            </attributes>
> > >          </meta_attributes>
> > >          <primitive id="resource_ipaddr_2" class="ocf" type="IPaddr"
> > > provider="heartbeat">
> > >            <instance_attributes id="resource_ipaddr_2_instance_attrs">
> > >              <attributes>
> > >  <nvpair id="d7945234-539b-4f86-9fd7-93501d1ff590" name="ip"
> > > value="66.199.245.204"/>
> > >                <nvpair id="c18f6b41-2798-42be-a5cb-470963ad3559"
> > name="nic"
> > > value="eth0"/>
> > >                <nvpair id="3112ab9f-d0e3-413e-84cd-76ecd39f51e7"
> > > name="cidr_netmask" value="255.255.255.224"/>
> > >              </attributes>
> > >            </instance_attributes>
> > >          </primitive>
> > >          <primitive id="resource_filesystem_2" class="ocf"
> > type="Filesystem"
> > > provider="heartbeat">
> > >            <instance_attributes
> > id="resource_filesystem_2_instance_attrs">
> > >              <attributes>
> > >                <nvpair id="db553cf1-1419-42d0-9a0b-caf30d2862a5"
> > > name="device"
> > > value="/dev/disk/by-uuid/81c3845e-c2f6-4cb0-a0cd-e00c074942fb"/>
> > >                <nvpair id="4f375334-b48f-4537-8351-ab5ae02eb351"
> > > name="directory" value="/vz"/>
> > >                <nvpair id="b1b5608e-0248-4a85-935d-b9957cdc044e"
> > > name="fstype" value="ext3"/>
> > >                <nvpair id="8a147356-405d-46dd-81ac-2cfa1401a988"
> > > name="options" value="_netdev,noatime"/>
> > >              </attributes>
> > >            </instance_attributes>
> > >          </primitive>
> > >          <primitive id="resource_vz_script_2" class="lsb" type="vz"
> > > provider="heartbeat">
> > >            <operations>
> > >              <op id="be25b095-6032-4890-b665-24b1860e3ab9" name="monitor"
> > > interval="15s" timeout="30s" start_delay="0s" on_fail="restart"/>
> > >              <op id="d715bd3a-851e-41cf-a836-93a3cfaa6a78" name="stop"
> > > timeout="200s"/>
> > >            </operations>
> > >          </primitive>
> > >        </group>
> > >      </resources>
> > >      <constraints>
> > >  <rsc_location id="location_slave_1" rsc="group_vz_1">
> > >          <rule id="prefered_location_slave_1" score="200">
> > >            <expression attribute="#uname"
> > > id="fd0fb0b6-3795-426d-966e-9da5fd24ff1a" operation="eq"
> > > value="node_slave"/>
> > >          </rule>
> > >        </rsc_location>
> > >        <rsc_location id="location_node3" rsc="group_vz_2">
> > >          <rule id="prefered_location_node3" score="600">
> > >            <expression attribute="#uname"
> > > id="72aaa5f0-c2f0-4038-afe3-4ce7cacb4acf" operation="eq" value="node3"/>
> > >          </rule>
> > >        </rsc_location>
> > >        <rsc_location id="location_slave_2" rsc="group_vz_2">
> > >   <rule id="prefered_location_slave_2" score="200">
> > >            <expression attribute="#uname"
> > > id="6cc8bf2f-9f52-4099-8b9e-85483973124c" operation="eq"
> > > value="node_slave"/>
> > >          </rule>
> > >        </rsc_location>
> > >        <rsc_location id="location_master" rsc="group_vz_1">
> > >          <rule id="prefered_location_master" score="600">
> > >            <expression attribute="#uname"
> > > id="2f8bfb3f-a6f7-4791-97d8-96d673e42ba6" operation="eq"
> > > value="node_master"/>
> > >          </rule>
> > >        </rsc_location>
> > >        <rsc_colocation id="colocation_vz" from="group_vz_1"
> > to="group_vz_2"
> > > score="-INFINITY"/>
> > >      </constraints>
> > >    </configuration>
> > >  </cib>
> > >
> > >
> > >
> > > Cheer's,
> > > Jaspal
> _______________________________________________
> Linux-HA mailing list
> [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
                                          
_________________________________________________________________
Hotmail: Free, trusted and rich email service.
http://clk.atdmt.com/GBL/go/171222984/direct/01/
_______________________________________________
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