Note: This bug is displayed in read-only format because
the product is no longer active in Red Hat Bugzilla.
RHEL Engineering is moving the tracking of its product development work on RHEL 6 through RHEL 9 to Red Hat Jira (issues.redhat.com). If you're a Red Hat customer, please continue to file support cases via the Red Hat customer portal. If you're not, please head to the "RHEL project" in Red Hat Jira and file new tickets here. Individual Bugzilla bugs in the statuses "NEW", "ASSIGNED", and "POST" are being migrated throughout September 2023. Bugs of Red Hat partners with an assigned Engineering Partner Manager (EPM) are migrated in late September as per pre-agreed dates. Bugs against components "kernel", "kernel-rt", and "kpatch" are only migrated if still in "NEW" or "ASSIGNED". If you cannot log in to RH Jira, please consult article #7032570. That failing, please send an e-mail to the RH Jira admins at rh-issues@redhat.com to troubleshoot your issue as a user management inquiry. The email creates a ServiceNow ticket with Red Hat. Individual Bugzilla bugs that are migrated will be moved to status "CLOSED", resolution "MIGRATED", and set with "MigratedToJIRA" in "Keywords". The link to the successor Jira issue will be found under "Links", have a little "two-footprint" icon next to it, and direct you to the "RHEL project" in Red Hat Jira (issue links are of type "https://issues.redhat.com/browse/RHEL-XXXX", where "X" is a digit). This same link will be available in a blue banner at the top of the page informing you that that bug has been migrated.
DescriptionMatthias Hensler
2013-01-06 07:50:37 UTC
Created attachment 673281[details]
Official upstream patch to fix rejoining
Description of problem:
We are running a six node cluster with corosync/pacemaker. A DDOS caused networkfailures causing the nodes to go berserk. After the problem was resolved one node refused to rejoin the cluster with "warning: cib_peer_callback: Discarding cib_apply_diff message (2810) from genesis: not in our membership".
That issue is fixed upstream, this patch is missing in current RHEL release.
Version-Release number of selected component (if applicable):
1.1.7-6.el6.x86_64
How reproducible:
Difficult. Although I did see the problem earlier in testing, I was not able to reprdouce it until it reoccured tonight. Several nodes leaving and rejoining the cluster because of network problems might be enough to trigger this.
Steps to Reproduce:
1. simulate DDOS by randomizing dropping networktraffic on all rings connected to the cluster
2. stop DDOS
3. check if all nodes rejoin
Actual results:
Out of 6 only 5 nodes proper rejoined. The 6th node seemed to rejoin (all other 5 nodes reported the 6th node joined and working), but refuses to accept any updates from its neighbours.
Log is full with
Jan 6 06:54:58 tim cib[44400]: warning: cib_peer_callback: Discarding cib_apply_diff message (2803) from genesis: not in our membership
Jan 6 06:54:58 tim cib[44400]: warning: cib_peer_callback: Discarding cib_apply_diff message (2804) from genesis: not in our membership
Jan 6 06:54:58 tim cib[44400]: warning: cib_peer_callback: Discarding cib_apply_diff message (2805) from genesis: not in our membership
Jan 6 06:54:58 tim cib[44400]: warning: cib_peer_callback: Discarding cib_apply_diff message (2806) from genesis: not in our membership
Expected results:
Node should properly rejoin
Additional info:
The issue is with pacemaker here. Corosync itself is fine, since "corosync-cfgtool -s" reports all rings up on every node and "corosync-objctl | grep member" lists all nodes joined on every other node.
Pacemaker lists all nodes operational on 5 out of 6 nodes, while the 6th node refuses to show any updates or accept any commands. Restarting pacemaker/corosync on that 6th node does not resolve the issue. It might be possible to resolve it by restarting all 6 nodes, but since they are productive I did not try that.
The problem seems to be known upstream and was discussed on the mailinglist. A summary can be found in this blogpost: http://www.tomas.cat/blog/en/troubleshooting-pacemaker-discarding-cibapplydiff-message-xxx-server2-not-our-membership
The fix described there, was to update to pacemaker 1.1.8. I did not try that, but instead recompiled the official pacemaker.src.rpm from RHEL6 only including the additional patch from https://github.com/ClusterLabs/pacemaker/commit/03f6105592281901cc10550b8ad19af4beb5f72f (also attached to this bug).
So far I only installed the recompiled pacemaker package on the affected 6th node. Starting pacemaker then immediately solved the issue.
Conclusion: I would strongly suggest to include the attached patch to the offical RHEL pacemaker package as soon as possible.