Login | Register For Free | Help
Search for: (Advanced)

Mailing List Archive: Linux-HA: Users

Q on http://clusterlabs.org/wiki/FAQ#Resource_is_Too_Active

 

 

Linux-HA users RSS feed   Index | Next | Previous | View Threaded


Ulrich.Windl at rz

Dec 15, 2011, 2:19 AM

Post #1 of 2 (303 views)
Permalink
Q on http://clusterlabs.org/wiki/FAQ#Resource_is_Too_Active

Hi!

I have a problem with some client-server software (I don't want to name it here) where client and server both need an etry for inetd (xinetd). It's also possible that client and server are running on one machine.

For a cluster solution I set up the server to run on one node only (using -INFINITY location restrictions for the other nodes). I've added a RA that diesables the indetd service when the server is to be down, and enabled it when the server is to be up.

Summary of a longer story: Even when the resource is never intended to run anywhere else, the cluster checks the indetd service through the RA on every node, and then complains:

Dec 15 09:10:55 h03 pengine: [15876]: WARN: See http://clusterlabs.org/wiki/FAQ#Resource_is_Too_Active for more information.

I'm afraid this might cause unintended fencing at some time.

Are there any rather clean solutions for this problem?

I was thinking to send the node's name to the RA that check's the inetd service, and to make him lie about the state on the other nodes, but I think that is terrible solution.

Cool ideas?

Regards,
Ulrich


_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems


dominik.klein at googlemail

Dec 16, 2011, 3:34 AM

Post #2 of 2 (279 views)
Permalink
Re: Q on http://clusterlabs.org/wiki/FAQ#Resource_is_Too_Active [In reply to]

On 12/15/2011 11:19 AM, Ulrich Windl wrote:
> Hi!
>
> I have a problem with some client-server software (I don't want to
> name it here) where client and server both need an etry for inetd
> (xinetd). It's also possible that client and server are running on
> one machine.
>
> For a cluster solution I set up the server to run on one node only
> (using -INFINITY location restrictions for the other nodes). I've
> added a RA that diesables the indetd service when the server is to be
> down, and enabled it when the server is to be up.
>
> Summary of a longer story: Even when the resource is never intended
> to run anywhere else, the cluster checks the indetd service through
> the RA on every node, and then complains:
>
> Dec 15 09:10:55 h03 pengine: [15876]: WARN: See
> http://clusterlabs.org/wiki/FAQ#Resource_is_Too_Active for more
> information.

Then your RA is wrong. It should not say that things are running when
they are not.

Regards
Dominik

> I'm afraid this might cause unintended fencing at some time.
>
> Are there any rather clean solutions for this problem?
>
> I was thinking to send the node's name to the RA that check's the
> inetd service, and to make him lie about the state on the other
> nodes, but I think that is terrible solution.
>
> Cool ideas?
>
> Regards, Ulrich
>
>
> _______________________________________________ Linux-HA mailing
> list Linux-HA [at] lists
> http://lists.linux-ha.org/mailman/listinfo/linux-ha See also:
> http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Linux-HA users RSS feed   Index | Next | Previous | View Threaded
 
 


Interested in having your list archived? Contact Gossamer Threads
 
  Web Applications & Managed Hosting Powered by Gossamer Threads Inc.