On 6/13/07, Jaime Medrano <[EMAIL PROTECTED]> wrote:
Hi.
I'm having problems when a one of the nodes changes its hostname (uname -n).
In first place, if the hostname is changed the cib doesn't get updated so
the '#uname' attribute is still seen as the old one in all the nodes of the
cluster. This is a minor problem, since the cluster keeps working as if the
hostname wasn't been changed.
The major problem arises when the node with the hostname changed reboots (a
heartbeat reload causes the same effect).
When this happens, that node can't see the others and no one can see it
either. In the logs there are messages like these:
Jun 13 16:25:44 p8100 cib: [19345]: WARN:
cib_peer_callback:callbacks.cDiscarding cib_slave_all message (84)
from i300b: not in our membership
Jun 13 16:25:44 p8100 cib: [19345]: WARN:
cib_peer_callback:callbacks.cDiscarding cib_apply_diff message (86)
from i300b: not in our membership
Jun 13 16:25:44 p8100 cib: [19345]: WARN:
cib_peer_callback:callbacks.cDiscarding cib_replace message (88) from
i300b: not in our membership
Jun 13 16:25:44 p8100 cib: [19345]: WARN:
cib_peer_callback:callbacks.cDiscarding cib_apply_diff message (8a)
from i300b: not in our membership
Jun 13 16:25:44 p8100 cib: [19345]: WARN:
cib_peer_callback:callbacks.cDiscarding cib_apply_diff message (8b)
from i300b: not in our membership
Jun 13 16:25:44 p8100 cib: [19345]: WARN:
cib_peer_callback:callbacks.cDiscarding cib_apply_diff message (8c)
from i300b: not in our membership
Jun 13 16:25:45 p8100 cib: [19345]: WARN:
cib_peer_callback:callbacks.cDiscarding cib_apply_diff message (8f)
from i300b: not in our membership
Jun 13 16:25:45 p8100 cib: [19345]: WARN:
cib_peer_callback:callbacks.cDiscarding cib_apply_diff message (90)
from i300b: not in our membership
Jun 13 16:25:46 p8100 cib: [19345]: WARN:
cib_peer_callback:callbacks.cDiscarding cib_apply_diff message (92)
from i300b: not in our membership
Jun 13 16:25:46 p8100 cib: [19345]: WARN:
cib_peer_callback:callbacks.cDiscarding cib_apply_diff message (93)
from i300b: not in our membership
(i300b is the new hostname)
Shouldn't be the uuid used instead the hostname in every places to avoid
this kind of problems?
it wont help much if the machine isn't reported as being part of the
cluster which is what the CIB is complaining about here
Regards,
Jaime Medrano.
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
