[Pacemaker] Problems with IPAddress2 type Resources causing SPLIT-BRAIN

[Димитър Бойн]

Thanks, Andrew!

I gave up the "eth0:0" idea.
The IP Resource seems to work with only specifying "eth0" too.

One "last" question :)
- when I use "ifconfig" I do not see any new alias created - though the resource is active as I can ssh to the box reported as hosting it.

What is the command/tool that I can use to identify I have active this Resource enabled IP?

Many thanks for your sincere support!

./Dimitar Boyn

-----Original Message-----
From: Andrew Beekhof [mailto:andrew at beekhof.net] 
Sent: Sunday, August 02, 2009 11:45 PM
To: pacemaker at clusterlabs.org
Subject: Re: [Pacemaker] Problems with IPAddress2 type Resources causing SPLIT-BRAIN

On Mon, Aug 3, 2009 at 1:07 AM, Димитър Бойн<dboyn at postpath.com> wrote:
> If I understand the question correctly no I am not.
> - i.e. I am attaching the resource IP to eth0:0 as visible in my example.
> I have a "management IP" on eht0.
>
> I am also getting a lot of the following on every node with active eth0:0:
> "
> Aug  2 16:02:16 ppst1node001 lrmd: [4980]: info: RA output: (ip-c001drbd10a:monitor:stderr) eth0:0: warning: name may be invalid
> Aug  2 16:02:58 ppst1node001 last message repeated 2 times
> Aug  2 16:04:01 ppst1node001 last message repeated 3 times
> Aug  2 16:05:04 ppst1node001 last message repeated 3 times
> "
>
> Is this an unsupported way of using the Resource?

Clearly not since the RA is complaining.

> How do I keep a constant IP for monitoring the node rather than the service if I do not use alias IP?

You don't.
Node monitoring happens independently of the resources.
There is no need to specify it manually


> Thanks!
> ./Dimitar
>
> -----Original Message-----
> From: Andrew Beekhof [mailto:andrew at beekhof.net]
> Sent: Sunday, August 02, 2009 3:19 PM
> To: pacemaker at clusterlabs.org
> Subject: Re: [Pacemaker] Problems with IPAddress2 type Resources causing SPLIT-BRAIN
>
> you're not creating resources for the nodes real ip addresses are you?
>
> On Sun, Aug 2, 2009 at 10:39 PM, Димитър Бойн<dboyn at postpath.com> wrote:
>> Hi,
>>
>> Has anybody experienced similar problem:
>>
>> IP address type Resource is still active while declaired stopped and the
>> cluster brings it up on a second node – the switches fence the node hosting
>> originally the Resouce and the cluster ends in SPLIT BRAIN situation?
>>
>> My resource is:
>>
>> „
>>
>> primitive ip-drbd00a ocf:heartbeat:IPaddr2 \
>>
>>         params ip="192.168.0.15" nic="eth0:0" broadcast="192.168.0.127"
>> cidr_netmask="25" \
>>
>>         op monitor interval="21s" timeout="5s" \
>>
>>         meta target-role="Started"
>>
>> „
>>
>>
>>
>> Аnée ideas how to avoid ?
>>
>>
>>
>> Thanks!
>>
>>
>>
>> ./Dimitar Boyn
>>
>> _______________________________________________
>> Pacemaker mailing list
>> Pacemaker at oss.clusterlabs.org
>> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>>
>>
>
> _______________________________________________
> Pacemaker mailing list
> Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> _______________________________________________
> Pacemaker mailing list
> Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>

_______________________________________________
Pacemaker mailing list
Pacemaker at oss.clusterlabs.org
http://oss.clusterlabs.org/mailman/listinfo/pacemaker