[Linux-HA] Looking for stonith documentation

2010-03-28 Thread mike
Hi All,

I'm looking for a good document that can take someoen very new to 
LinuxHA (me) and clearly explain how stonith is implemented and how to 
configure it. Most of what I've found so far is very cryptic and frankly 
dose not explain how to configure stonith in a working cluster. For my 
example, I have 3 working IP Takeover clusters running httpd.  What I 
want to do should be simple I would think. If there is a takeover, kill 
the other node. I have a command  that works very well in the 
environment I am using LinuxHA but I can't find any documentation that 
helps me understand how stonith is implemented and configured. I need to 
understand this first.

Can someone point me to some good documentation on stonith? Where do I 
start?
___
Linux-HA mailing list
Linux-HA@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems


Re: [Linux-HA] Looking for stonith documentation

2010-03-28 Thread Tim Serong
On 3/29/2010 at 10:24 AM, mike mgbut...@nbnet.nb.ca wrote: 
 Hi All, 
  
 I'm looking for a good document that can take someoen very new to  
 LinuxHA (me) and clearly explain how stonith is implemented and how to  
 configure it. Most of what I've found so far is very cryptic and frankly  
 dose not explain how to configure stonith in a working cluster. For my  
 example, I have 3 working IP Takeover clusters running httpd.  What I  
 want to do should be simple I would think. If there is a takeover, kill  
 the other node. I have a command  that works very well in the  
 environment I am using LinuxHA but I can't find any documentation that  
 helps me understand how stonith is implemented and configured. I need to  
 understand this first. 
  
 Can someone point me to some good documentation on stonith? Where do I  
 start?

Try http://www.clusterlabs.org/doc/crm_fencing.html

Also various links at http://www.clusterlabs.org/wiki/Documentation
(in particular Clusters from Scratch and Configuration Explained).

Maybe http://ourobengr.com/ha (but that's more about what can go wrong
with STONITH in a two node cluster, than about how to configure STONITH).

Note that whether or not STONITH occurs for a given failure depends on
how you've configured your resources.  It'll always trigger if a node
is (apparently) completely dead, or (by default) if a resource stop op
fails.  If a resource can't start, or is running then fails, STONITH
won't occur unless you set the on-fail property of that operation to
fence.  You should find more detail about this in the Configuration
Explained document.

HTH,

Tim


-- 
Tim Serong tser...@novell.com
Senior Clustering Engineer, OPS Engineering, Novell Inc.



___
Linux-HA mailing list
Linux-HA@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems