Re: [Linux-HA] I've got two groups, taking down group2 causes group1 to do down and then up again.

Sun, 17 Jun 2007 11:37:02 -0700 

Well, its me again, talking to myself.
To set "ordered" to false causes all the resources in the group to start in no order at all. 
Thats not what I wanted.
Anyway - the problem disappeared when I set "ordered" to true instead.
But when I remove all files in /var/lib/heartbeat/crm/ and start up the cluster 
again and lets heartbeat create a base cib.xml and then adding my groups
with their resources, and also sets resource_stickiness to INFINITY
it works like a breeze again. I dont get it.
Maybe its me who have some gnomes in my head...

Anyway - I hope it will be the last post from me in this subject.

Regards
Claes

Claes Lindvall skrev:

I found a solution for my problem.
From the GUI I could set resource_stickiness to INFINITY.
And then I found an attribute named "ordered". If I set it to "false"
In a tutorial a found an explanation of the attribute "ordered":
ordered – boolean – defaults to TRUE
TRUE means the group physical ordering implies start-after
ordering constraints
FALSE means no such start-after ordering is implied

If there is another solution that is better, please let me know.

Regards
Claes Lindvall


Claes Lindvall skrev:
Let me further explain what I do, it seems to me that I somewhere is missing a detail, or thinking wrong. 

I have 2 nodes.
I do everythings from the GUI.
I create two groups.
Then I create a lsb resource apache2 and make it belong to group1
Then I create a lsb resource ntp and make it belong to group2.
I create a "place" and lets group1 be associated with one of the nodes by the expression #uname eq <one_of_my_nodes> 
Same for group2, but towards the other node.
I remove "target_role" from my resources and addresses all commands (started, stopped) to the groups only. 
I fire it up and everything is running perfect.
Now (here it comes) when I am taking down group2 , group1 is going down also, and then group2. 
When group2 is down - group1 goes up again.
Im used to HPUX MC/ServiceGuard and are mapping my thoughts directly to heartbeat, and that is not allways right to do. Am I thinking completly wrong, is this normal behavior, and what do I 
do to achive my goal - to let my groups be indipendet from each other.

If I work directly with resources (eg no groups) this does not happen.
As soon as I work with groups these problems (?) occurs.
By the way - thanks for a very dynamic and nice piece of software. Im impressed. 

Kindly Regards
Claes Lindvall



Claes Lindvall skrev:
Some weeks ago I posted a query about my lab-cluster - when Im taking down pkg1 (one of the groups) 
pkg2 is also taken down and then up again.
I got a tip about using the default_resource_stickiness. And that caused group1 to do down and then up again without changing node. And that is better, but not what I want. 
I want to take down group2 without disturb group1.
But I guess something more needs to be done in my (attached) config (cibadmin -Ql) 

Im running SLES 10
heartbeat-pils-2.0.7-1.2
heartbeat-ldirectord-2.0.7-1.2
yast2-heartbeat-2.13.10-1.3
heartbeat-stonith-2.0.7-1.2
heartbeat-cmpi-2.0.7-1.2
heartbeat-2.0.7-1.2

Thanks,
Claes Lindvall
------------------------------------------------------------------------
<cib admin_epoch="0" have_quorum="true" num_peers="2" cib_feature_revision="1.3" ccm_transition="2" generated="true" dc_uuid="19469e0b-4e6f-4688-b827-863b2d2bec66" epoch="2" num_updates="295"> 
<configuration>
<crm_config>
<cluster_property_set id="cib-bootstrap-options">
<attributes>
<nvpair name="last-lrm-refresh" id="cib-bootstrap-options-last-lrm-refresh" value="1181894748"/> <nvpair id="cib-bootstrap-options-default_resource_stickiness" name="default_resource_stickiness" value="INFINITY"/> <nvpair id="cib-bootstrap-options-stonith_enabled" name="stonith_enabled" value="True"/> <nvpair name="no_quorum_policy" id="cib-bootstrap-options-no_quorum_policy" value="stop"/> 
</attributes>
</cluster_property_set>
</crm_config>
<nodes>
<node id="19469e0b-4e6f-4688-b827-863b2d2bec66" uname="lx11" type="normal"/> <node id="0c6792aa-4a4f-467a-964e-15efa08bd7b5" uname="lx10" type="normal"/> 
</nodes>
<resources>
<group resource_stickiness="INFINITY" id="pkg1">
<instance_attributes id="pkg1_instance_attrs">
<attributes>
<nvpair name="target_role" id="pkg1_target_role" value="started"/>
</attributes>
</instance_attributes>
<primitive class="lsb" type="apache2" provider="heartbeat" resource_stickiness="#default" id="pkg1_web"> 
<instance_attributes id="pkg1_web_instance_attrs">
<attributes/>
</instance_attributes>
</primitive>
</group>
<group resource_stickiness="INFINITY" id="pkg2">
<instance_attributes id="pkg2_instance_attrs">
<attributes>
<nvpair name="target_role" id="pkg2_target_role" value="stopped"/>
</attributes>
</instance_attributes>
<primitive class="lsb" type="ntp" provider="heartbeat" id="pkg2_ntp" resource_stickiness="#default"> 
<instance_attributes id="pkg2_ntp_instance_attrs">
<attributes/>
</instance_attributes>
</primitive>
</group>
</resources>
<constraints/>
</configuration>
<status>
<node_state uname="lx11" crmd="online" shutdown="0" in_ccm="true" ha="active" join="member" expected="member" id="19469e0b-4e6f-4688-b827-863b2d2bec66"> 
<lrm id="19469e0b-4e6f-4688-b827-863b2d2bec66">
<lrm_resources>
<lrm_resource type="apache2" class="lsb" provider="heartbeat" id="pkg1_web"> <lrm_rsc_op id="pkg1_web_monitor_0" operation="monitor" crm-debug-origin="do_update_resource" transition_key="51:13e8a2f0-7188-41a0-9c60-7309007bac0f" transition_magic="4:7;51:13e8a2f0-7188-41a0-9c60-7309007bac0f" call_id="19" crm_feature_set="1.0.6" rc_code="7" op_status="4" interval="0" op_digest="e1c46569e84224c134048167b5c96b36"/> <lrm_rsc_op operation="start" crm_feature_set="1.0.6" rc_code="0" op_status="0" interval="0" op_digest="e1c46569e84224c134048167b5c96b36" id="pkg1_web_start_0" transition_key="62:13e8a2f0-7188-41a0-9c60-7309007bac0f" transition_magic="0:0;62:13e8a2f0-7188-41a0-9c60-7309007bac0f" call_id="27"/> <lrm_rsc_op id="pkg1_web_stop_0" operation="stop" crm-debug-origin="do_update_resource" transition_key="62:13e8a2f0-7188-41a0-9c60-7309007bac0f" transition_magic="0:0;62:13e8a2f0-7188-41a0-9c60-7309007bac0f" call_id="25" crm_feature_set="1.0.6" rc_code="0" op_status="0" interval="0" op_digest="e1c46569e84224c134048167b5c96b36"/> 
</lrm_resource>
<lrm_resource type="ntp" class="lsb" provider="heartbeat" id="pkg2_ntp"> <lrm_rsc_op id="pkg2_ntp_monitor_0" operation="monitor" crm-debug-origin="do_update_resource" transition_key="52:13e8a2f0-7188-41a0-9c60-7309007bac0f" transition_magic="4:7;52:13e8a2f0-7188-41a0-9c60-7309007bac0f" call_id="20" crm_feature_set="1.0.6" rc_code="7" op_status="4" interval="0" op_digest="f2317cad3d54cec5d7d7aa7d0bf35cf8"/> <lrm_rsc_op operation="start" crm_feature_set="1.0.6" rc_code="0" op_status="0" interval="0" op_digest="f2317cad3d54cec5d7d7aa7d0bf35cf8" id="pkg2_ntp_start_0" transition_key="59:13e8a2f0-7188-41a0-9c60-7309007bac0f" transition_magic="0:0;59:13e8a2f0-7188-41a0-9c60-7309007bac0f" call_id="23"/> <lrm_rsc_op operation="stop" crm_feature_set="1.0.6" rc_code="0" op_status="0" interval="0" op_digest="f2317cad3d54cec5d7d7aa7d0bf35cf8" id="pkg2_ntp_stop_0" transition_key="62:13e8a2f0-7188-41a0-9c60-7309007bac0f" transition_magic="0:0;62:13e8a2f0-7188-41a0-9c60-7309007bac0f" call_id="26"/> 
</lrm_resource>
</lrm_resources>
</lrm>
<transient_attributes id="19469e0b-4e6f-4688-b827-863b2d2bec66">
<instance_attributes id="status-19469e0b-4e6f-4688-b827-863b2d2bec66">
<attributes>
<nvpair id="status-19469e0b-4e6f-4688-b827-863b2d2bec66-probe_complete" name="probe_complete" value="true"/> 
</attributes>
</instance_attributes>
</transient_attributes>
</node_state>
<node_state uname="lx10" crmd="online" in_ccm="true" ha="active" join="member" expected="member" shutdown="0" id="0c6792aa-4a4f-467a-964e-15efa08bd7b5"> 
<lrm id="0c6792aa-4a4f-467a-964e-15efa08bd7b5">
<lrm_resources>
<lrm_resource type="apache2" class="lsb" provider="heartbeat" id="pkg1_web"> <lrm_rsc_op id="pkg1_web_monitor_0" operation="monitor" crm-debug-origin="do_update_resource" transition_key="51:13e8a2f0-7188-41a0-9c60-7309007bac0f" transition_magic="4:7;51:13e8a2f0-7188-41a0-9c60-7309007bac0f" call_id="41" crm_feature_set="1.0.6" rc_code="7" op_status="4" interval="0" op_digest="e1c46569e84224c134048167b5c96b36"/> <lrm_rsc_op operation="start" crm_feature_set="1.0.6" rc_code="0" op_status="0" interval="0" op_digest="e1c46569e84224c134048167b5c96b36" id="pkg1_web_start_0" transition_key="59:13e8a2f0-7188-41a0-9c60-7309007bac0f" transition_magic="0:0;59:13e8a2f0-7188-41a0-9c60-7309007bac0f" call_id="53"/> <lrm_rsc_op operation="stop" crm_feature_set="1.0.6" rc_code="0" op_status="0" interval="0" op_digest="e1c46569e84224c134048167b5c96b36" id="pkg1_web_stop_0" transition_key="60:13e8a2f0-7188-41a0-9c60-7309007bac0f" transition_magic="0:0;60:13e8a2f0-7188-41a0-9c60-7309007bac0f" call_id="54"/> 
</lrm_resource>
<lrm_resource id="pkg2_ntp" type="ntp" class="lsb" provider="heartbeat"> <lrm_rsc_op id="pkg2_ntp_monitor_0" operation="monitor" crm-debug-origin="do_update_resource" transition_key="52:13e8a2f0-7188-41a0-9c60-7309007bac0f" transition_magic="4:7;52:13e8a2f0-7188-41a0-9c60-7309007bac0f" call_id="42" crm_feature_set="1.0.6" rc_code="7" op_status="4" interval="0" op_digest="f2317cad3d54cec5d7d7aa7d0bf35cf8"/> 
</lrm_resource>
</lrm_resources>
</lrm>
<transient_attributes id="0c6792aa-4a4f-467a-964e-15efa08bd7b5">
<instance_attributes id="status-0c6792aa-4a4f-467a-964e-15efa08bd7b5">
<attributes>
<nvpair id="status-0c6792aa-4a4f-467a-964e-15efa08bd7b5-probe_complete" name="probe_complete" value="true"/> 
</attributes>
</instance_attributes>
</transient_attributes>
</node_state>
</status>
</cib>
------------------------------------------------------------------------ 

_______________________________________________
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




_______________________________________________
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