On 2006-01-24T17:23:39, Andrew Beekhof <[EMAIL PROTECTED]> wrote:

> > # cibadmin -Q | grep transition_idle_timeout
> >              <nvpair id="transition_idle_timeout" 
> > name="transition_idle_timeout" value="120s"/>
> > # cibadmin -Q | sed 
> > '/transition_idle_timeout/s/value="120s"/value="200s"/'| cibadmin -U -p
> > Call cib_update failed (-25): The update was empty
> > <null>
> > #

> If you add a few -V options you'll be able to see what cibadmin thinks
> you're sending it.
> Also, in 2.0.3/CVS the option you need is -M
> 
> -U would require all the parent nodes (with id's where appropriate)
> which is not as simple to construct and the reason -M was added.

Well he IS sending in the full parent nodes. He's basically feeding back
an entire CIB with just that change, and I think that includes feeding
back various state data etc. Mayb'e that's screwing up stuff?

-M and feeding back just the changed timeout line would probably make it
a lot easier to debug.


Sincerely,
    Lars Marowsky-Brée

-- 
High Availability & Clustering
SUSE Labs, Research and Development
SUSE LINUX Products GmbH - A Novell Business     -- Charles Darwin
"Ignorance more frequently begets confidence than does knowledge"

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

Reply via email to