Red Hat Bugzilla – Bug 195345
device-mapper goes crazy after a cmirror relocation
Last modified: 2010-04-27 10:54:39 EDT
Description of problem:
I realize that we are about to get a new build which will hopefully fix a lot of
issues with cmirrors, but I figured this issue should still be documented and
checked out once we get that fix.
I was attempting to verify the fix for bz 193596, by doing a mirror server
relocaion during a mirror creation, and that caused device-mapper to spit out
thousands and thousands of the following messages on one of the cluster nodes
along with causing the lvcreate cmd to hang.
Jun 14 09:58:02 taft-02 kernel: device-mapper: Unable to convert
nodeid_to_ipaddr in _consult_server
Jun 14 09:58:02 taft-02 kernel: device-mapper: Clustered mirror retried requests
:: 51357152 of 51358401 (17%)
Jun 14 09:58:02 taft-02 kernel: device-mapper: No address list for nodeid 0
Version-Release number of selected component (if applicable):
[root@taft-02 ~]# rpm -q cmirror
[root@taft-02 ~]# rpm -q cmirror-kernel
I've changed these to be 'debug' prints. Depending on what your logging level is set at, you may still
see these. Besides just the prints, I've worked to compensate for when these messages happen... so
you shouldn't get that many of them.
This message could be removed if we wanted, but obviously, a hang would be bad.
Assuming this VERIFIED fix got released. Closing.
Reopen if it's not yet resolved.