andrew at beekhof.net
Mon Sep 6 12:21:00 MDT 2004
On Sep 6, 2004, at 3:05 PM, Lars Marowsky-Bree wrote:
> RID should be imported from lrm_api.h instead of being locally defined
> (and wrongly so).
> (And probably be renamed to LRM_RSC_ID_LEN or something more sane.)
> The timeouts are not yet pulled in from the resources definition in the
nod - i believe i even mentioned this at one point :)
> For monitor(), is target_rc == CHANGED really the same as target_rc ==
No. But the assumption i had that we know after the first callback
that it failed, receiving followups when nothing has changed didnt seem
to add anything. unless we changed DC, but even then we would have
done a query for that information.
> We need to pass in the monitoring depth (and then the LRM needs to put
> that into the OCF_MONITOR_DEPTH variable for OCF agents), and handle
> monitors of different depths.
right, not done yet.
> Given that pulling the version number of the RA out inside the LRM
> mean that lrmd needs to handle libxml2, I think we should extract this
> ourselves in the meta-data operation.
that would seem to make sense.
> (I'm not trying to annoy anyone, just trying to get a feel for the
> various pieces of code and pointing out what I find or don't understand
> yet ;-)
about time you went over my code! ;)
> Lars Marowsky-Brée <lmb at suse.de>
> High Availability & Clustering \\\ ///
> SUSE Labs, Research and Development \honk/
> SUSE LINUX AG - A Novell company \\//
> Linux-HA-Dev: Linux-HA-Dev at lists.linux-ha.org
> Home Page: http://linux-ha.org/
"Ooo Ahhh, Glenn McRath" - TISM
More information about the Linux-HA-Dev