Mailing List Archive: Linux-HA: Dev

Problems with configuration

Index | Next | Previous | View Threaded

sid at kongregate

Jun 20, 2008, 9:58 AM

Post #1 of 2 (254 views)
Permalink

Problems with configuration

Hi,

We are trying to use heartbeat2 to manage several memcached servers. There
are three requirements

1. Run one and only one memcached server on 4 of the 5 available servers
2. Once a memcached server starts on a node, leave it running until the
node is shut down.
3. Associate a virtual IP address with each active memcached process.

While the first two requirements are easy to implement using clones, we have
not found a way to clone the ipaddress OCF resource while assigning unique
addresses to different hosts for third requirement. Nor does there seem to
be a way to define the ipaddress service as a separate resource and then
co-locate it with a clone.

We have created resource groups that include the ipaddress OCF resource and
our memcached OCF resource. But we have not been able to find a set of
contraints that meet our requirements. We have tried,

- Set symetric cluster to true, with no contraints defined. All three
resources were started on the same server.
- Create co-location contraints that limit the groups to not run on the
same node as other groups. This configuration fulfils requirement 1, but for
some reason heartbeat restarts/moves resource groups when a new resource is
added. Note that we have Default Resource Stickiness set to INFINITY.
- We then turned off symetric cluster and created locations that allowed
the resources to run on all servers. This generated a cib.xml error warning
and none of the resources started.

Does anyone have a suggestion for the correct settings/contraints to fulfill
our requirements? I have read through the documentation on the web site as
well at the DTD for the cib.xml. Are there any other resources that describe
how to configure the software?

Sid Stuart

lmb at suse

Jun 20, 2008, 1:16 PM

Post #2 of 2 (234 views)
Permalink

Re: Problems with configuration [In reply to]

On 2008-06-20T09:58:16, Sid Stuart <sid[at]kongregate.com> wrote:

>
> 1. Run one and only one memcached server on 4 of the 5 available servers
> 2. Once a memcached server starts on a node, leave it running until the
> node is shut down.
> 3. Associate a virtual IP address with each active memcached process.

You may want to have a way for the IPaddr2 clone to not just go into
cluster-alias mode, but into an "additive" mode, where the RA would
simply add the clone number to the base IP and then configure that as a
single IP (if you see what I'm saying).

Regards,
Lars

--
Teamlead Kernel, SuSE Labs, Research and Development
SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg)
"Experience is the name everyone gives to their mistakes." -- Oscar Wilde

_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/

Index | Next | Previous | View Threaded

Interested in having your list archived? Contact lists@gossamer-threads.com