Red Hat Bugzilla – Bug 1312092
crmd can crash after unexpected remote connection takeover
Last modified: 2016-05-10 19:52:58 EDT
Description of problem: An attempted connection to a Pacemaker Remote node that already is integrated into a cluster will disconnect the node from the cluster, and the cluster host's crmd may print garbled log messages and/or crash.
Version-Release number of selected component (if applicable):
How reproducible: easily/consistently
Steps to Reproduce:
1. Configure a Pacemaker cluster including a Pacemaker Remote node.
2. Start the cluster and wait for the remote node to integrate.
3. From any machine with access to the Pacemaker Remote port, run "nc -t <nodename> 3121".
Actual results: Remote node will be disconnected from the cluster, its monitor operation will fail, and the cluster node hosting its connection may have crmd crash or print log messages with random characters.
Expected results: The attempted connection fails, and the cluster is unaffected.
This was fixed upstream as of 5ec24a2, and the necessary patches were included in the latest build for #1252206. This BZ was added for tracking and QA purposes.
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.