[Pacemaker] getting started

[Frank DiMeo]

Resending without the mangled grammar--

 

Thanks for your response,  some comments:

 

1)       So, if we cannot load a persistent configuration after a restart of the all nodes in a cluster, how does one automatically configure a (rebooted) unattended cluster with a known set of resources?  Is it expected that a person will reconfigure a cluster if all nodes in the cluster were to reboot?

 

2)       If I'm trying to reduce the DC election time during a complete cluster reboot, so how can I load the dc-deadtime before a DC is elected?

 

3)       Thanks, that's was the problem

 

4)       I'm sure your method will work, but the syntax that I used is right out of the documentation for pacemaker (example 6.16 of Pacemaker 1.0, Configuration explained"), so I don't see what's wrong with it.  Regardless, I'll use your method.

 

-Frank

 

 

________________________________

From: Frank DiMeo [mailto:Frank.DiMeo at bigbandnet.com] 
Sent: Wednesday, November 18, 2009 4:28 PM
To: pacemaker at oss.clusterlabs.org
Subject: Re: [Pacemaker] getting started

 

Thanks for you response,  some comments:

 

5)   So, if we cannot load a persistent configuration after a restart of the all nodes in the cluster, how does one automatically configure an (rebooted) unattended cluster with a known set of resources?  Is it expected that a person will reconfigure a cluster if it were to all nodes in the cluster were to reboot?

 

6)   If I'm trying to reduce the DC election time during a complete cluster reboot, so how can I load the dc-deadtime before a DC is elected?

 

7)   Thanks, that's was the problem

 

8)   I'm sure your method will work, but the syntax that I used is right out of the documentation for pacemaker (example 6.16 of Pacemaker 1.0, Configuration explained"), so I don't see what's wrong with it.  Regardless, I'll use your method.

 

-Frank

 

________________________________

From: hj lee [mailto:kerdosa at gmail.com] 
Sent: Wednesday, November 18, 2009 4:12 PM
To: pacemaker at oss.clusterlabs.org
Subject: Re: [Pacemaker] getting started

 

 

	1) Is there a way to start pacemaker and specify an xml file as the startup configuration?  Right now, I start openais/pacemaker, wait for DC election, then run cibadmin to load the file.  That's fine for now, but I want to automate this eventually.

 
I don't think it's possible.

 

	2) Is there any way to shorten the time for DC election?  I was hoping to be able to specify a cluster-delay shorter than 60s by loading an xml file at startup, and thereby shorten this election time.


crm configure property dc-deadtime=10s

 

	3) I can't seem to specify either an rsc_defaults or op_defaults section in an xml file.  When I try to specify rsc_defaults for instance, syslog shows:
	
	ov 18 08:57:57 ubuntu_1 cibadmin: [27658]: info: Invoked: cibadmin --replace --xml-file 2rcordered_1_sticky.xml
	Nov 18 08:57:57 ubuntu_1 cib: [27081]: ERROR: Extra element rsc_defaults in interleave
	Nov 18 08:57:57 ubuntu_1 cib: [27081]: ERROR: Element configuration failed to validate content
	Nov 18 08:57:57 ubuntu_1 cib: [27081]: ERROR: Element cib failed to validate content
	Nov 18 08:57:57 ubuntu_1 cib: [27081]: WARN: cib_perform_op: Updated CIB does not validate against pacemaker-1.0 schema/dtd
	Nov 18 08:57:57 ubuntu_1 cib: [27081]: WARN: cib_diff_notify: Update (client: cibadmin, call:2): 0.7.3 -> 3.29.1 (Update does not conform to the configured schema/DTD)

 
As the log suggested, the xml does not comform to pacemaker RelaxNG schema. If you look /usr/share/pacemaker/pacemaker-1.0.rng, the rsc_defaults requires element name "meta_attributes". So please change it to this: 

 <rsc_defaults>
      <meta_attributes id="testid">
      <nvpair id="rsc-default-1" name="resource-stickiness" value="100"/>
      </meta_attributes>
  </rsc_defaults>

	
	4) I'm trying to specify two collocated resources that don't have any startup dependency.  When I use the syntax:
	
	   <constraints>
	     <rsc_colocation id="coloc-1" score="INFINITY">
	       <resource_set id="collocated-set-1" sequential="false">
	         <resource_ref id="eyes"/>
	         <resource_ref id="clock"/>
	       </resource_set>
	     </rsc_colocation>
	   </constraints>

It does not seem correct. Please use crm configure command to configure the cluster rather than xml until you you become very famillar with it.
crm configure colocation coloc-1 inf: clock eyes
Isn't it much simpler?

 Thanks
hj

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clusterlabs.org/pipermail/pacemaker/attachments/20091118/ffcd2ddf/attachment-0002.htm>