On 2006-05-26T09:02:35, Andrew Beekhof <[EMAIL PROTECTED]> wrote:
> >notify: post for stop
> >notify active: xen-4 xen-3 xen-1
> >notify stop: xen-2
> >notify start: xen-4 xen-3 xen-2 xen-1
> do you have an example of this? nothing i have shows this behavior.
It's the same CIB/cluster I've sent you several times, so I don't think there's
much value in me sending it again ;-)
Here's a shortened extract of a cluster start + failed node:
17:14:26 xen-1 Filesystem[1754]: DEBUG: - notify: pre for start
17:14:26 xen-1 Filesystem[1754]: DEBUG: - notify active:
17:14:26 xen-1 Filesystem[1754]: DEBUG: - notify stop:
17:14:26 xen-1 Filesystem[1754]: DEBUG: - notify start: xen-4 xen-3 xen-2 xen-1
17:14:45 xen-1 Filesystem[1813]: DEBUG: - notify: post for start
17:14:45 xen-1 Filesystem[1813]: DEBUG: - notify active:
17:14:45 xen-1 Filesystem[1813]: DEBUG: - notify stop:
17:14:45 xen-1 Filesystem[1813]: DEBUG: - notify start: xen-4 xen-3 xen-2 xen-1
17:14:45 xen-1 Filesystem[1813]: DEBUG:: ignoring post-notify for start.
So, the initial start is right; all as expected.
Now xen-2 has met its untimely demise, and the cluster reacts:
17:17:02 xen-1 Filesystem[1941]: DEBUG: - notify: pre for stop
17:17:02 xen-1 Filesystem[1941]: DEBUG: - notify active: xen-4 xen-3 xen-1
17:17:02 xen-1 Filesystem[1941]: DEBUG: - notify stop: xen-2
17:17:02 xen-1 Filesystem[1941]: DEBUG: - notify start: xen-4 xen-3 xen-2 xen-1
17:17:02 xen-1 Filesystem[1941]: DEBUG:: ignoring pre-notify for stop.
17:17:05 xen-1 Filesystem[1965]: DEBUG: - notify: post for stop
17:17:05 xen-1 Filesystem[1965]: DEBUG: - notify active: xen-4 xen-3 xen-1
17:17:05 xen-1 Filesystem[1965]: DEBUG: - notify stop: xen-2
17:17:05 xen-1 Filesystem[1965]: DEBUG: - notify start: xen-4 xen-3 xen-2 xen-1
The stop notifications are already wrong. notify start doesn't fit
reality.
Now, sometime later, xen-2 reappears:
17:43:47 xen-1 Filesystem[2712]: DEBUG: - notify: pre for start
17:43:47 xen-1 Filesystem[2712]: DEBUG: - notify active: xen-4 xen-3 xen-1
17:43:47 xen-1 Filesystem[2712]: DEBUG: - notify stop: xen-2
17:43:47 xen-1 Filesystem[2712]: DEBUG: - notify start: xen-2
17:43:54 xen-1 Filesystem[2748]: DEBUG: - notify: post for start
17:43:54 xen-1 Filesystem[2748]: DEBUG: - notify active: xen-4 xen-3 xen-1 y
17:43:54 xen-1 Filesystem[2748]: DEBUG: - notify stop: xen-2 z
17:43:54 xen-1 Filesystem[2748]: DEBUG: - notify start: xen-2
17:43:54 xen-1 Filesystem[2748]: DEBUG:: ignoring post-notify for start.µ
Hrm. Same here, just this time around, it's the stop which doesn't fit.
I've taken the PE inputs and ran them through, but these don't show up
in the transition graph. Seems to be introduced later.
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/
