Bug 226700 - cman cluster needs restart when going from >=3 to 2 nodes and 2 to >= 3 nodes
cman cluster needs restart when going from >=3 to 2 nodes and 2 to >= 3 nodes
Product: Red Hat Enterprise Linux 5
Classification: Red Hat
Component: conga (Show other bugs)
All Linux
medium Severity medium
: ---
: ---
Assigned To: Ryan McCabe
Corey Marthaler
Depends On: 240508
  Show dependency treegraph
Reported: 2007-01-31 19:06 EST by Paul Kennedy
Modified: 2015-04-19 20:47 EDT (History)
7 users (show)

See Also:
Fixed In Version: RHSA-2007-0640
Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of:
Last Closed: 2007-11-07 10:37:00 EST
Type: ---
Regression: ---
Mount Type: ---
Documentation: ---
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---

Attachments (Terms of Use)

  None (edit)
Description Paul Kennedy 2007-01-31 19:06:59 EST
Description of problem:
Tried to delete a node and add it back into the cluster. Received this error
message when trying to add the node back in:

The following errors occurred:
    * A Ricci error occurred on tng3-4.lab.msp.redhat.com: ccs_tool failed
    * Unable to update the cluster node list for my_rh_cluster

Further investigation revealed this:
[root@tng3-3 ~]# cman_tool nodes
Node  Sts   Inc   Joined               Name
   1   X     12                        tng3-5.lab.msp.redhat.com
   2   M     12   2007-01-31 17:26:00  tng3-4.lab.msp.redhat.com
   3   M      4   2007-01-31 17:26:00  tng3-3.lab.msp.redhat.com

[root@tng3-4 ~]# cman_tool nodes
Node  Sts   Inc   Joined               Name
   1   X      8                        tng3-5.lab.msp.redhat.com
   2   M      4   2007-01-31 12:25:03  tng3-4.lab.msp.redhat.com
   3   M     12   2007-01-31 12:25:09  tng3-3.lab.msp.redhat.com

Version-Release number of selected component (if applicable):

How reproducible:
Create three-node cluster. Remove one node and add it back.

Steps to Reproduce:
Actual results:

Expected results:

Additional info:
Comment 1 Paul Kennedy 2007-01-31 19:13:28 EST
Tried adding different node. Same results.
Comment 2 Ryan McCabe 2007-02-01 10:03:36 EST
This looks like a bug somewhere deeper in the cluster stack. Could you try to
manually propagate the conf file with ccs_tool and see what the error message is?
Comment 4 Ryan McCabe 2007-02-01 11:14:10 EST
After I remove the node, then try to add it back manually, here's what I see:

[root@huey cluster]# ccs_tool update .cluster.conf 
Unable to open connection to dewey.lab.boston.redhat.com: Bad file descriptor

Failed to update config file.
[root@huey cluster]# cman_tool nodes
Node  Sts   Inc   Joined               Name
   1   M     12   2007-02-01 11:06:44  louey.lab.boston.redhat.com
   2   M      4   2007-02-01 11:06:44  huey.lab.boston.redhat.com
   3   X     12                        dewey.lab.boston.redhat.com

[root@dewey cluster]# pwd;ls -la
total 32
drwxr-xr-x  2 root root  4096 Feb  1 11:10 .
drwxr-xr-x 91 root root 12288 Feb  1 11:07 ..
-rw-r-----  1 root root   461 Feb  1 11:07 .cluster.conf

[root@dewey cluster]# ps auxww|egrep "[a]isexec|[c]cs|[c]man|[c]lurg";service
cman status
ccsd is stopped

Anyone have any insight into what's going wrong here?
Comment 5 Ryan McCabe 2007-02-01 11:22:16 EST
Sorry. About 5 minutes after I posted here, i realized that the cluster needs a
restart when going from 3->2 and 2->3 nodes. That's what's causing the problem.
Comment 6 Kiersten (Kerri) Anderson 2007-04-23 13:06:27 EDT
Fixing Product Name.  Cluster Suite was merged into Enterprise Linux for version
Comment 7 Ryan McCabe 2007-05-18 12:44:31 EDT
Marking this modified and depending on

That fix will remove the need for any special handling on the part of conga for
clusters going from > 2 nodes to 2 and from 2 to > 2.
Comment 10 errata-xmlrpc 2007-11-07 10:37:00 EST
An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on the solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.


Note You need to log in before you can comment on or make changes to this bug.