Werk #6358: Fixed stale services on cluster nodes
Component | Core & setup |
Title | Fixed stale services on cluster nodes |
Date | Jul 21, 2018 |
Checkmk Edition | Checkmk Raw (CRE) |
Checkmk Version | 1.5.0b10 1.6.0b1 |
Level | Prominent Change |
Class | Bug Fix |
Compatibility | Compatible - no manual interaction needed |
When using Check_MK clusters it could happen that some of the services on a node went stale and remained in this state. In this situation you can find multiple messages in the cmc.log when using the CEE:
2018-07-21 12:47:52 [5] [Check_MK helper 4644] started, commandline: /omd/sites/beta/bin/cmk --keepalive 2018-07-21 12:48:42 [2] [Check_MK helper 4644] ignoring check result for lxclu1 / Filesystem /boot: no such service 2018-07-21 12:48:43 [4] [Check_MK helper 4644] restarting dead helper 2018-07-21 12:48:43 [5] [Check_MK helper 4644] exited normally
The issue was triggered because the cluster nodes reported service check results for services which are assigned to the cluster using the "Clustered services" rule set.
The problem could only happen when services of one check type were assigned to the cluster and the cluster node. For example in case you have at least one Filesystem service assigned to the node and at least one Filesystem service assigned to the cluster.
This regression was introduced with 1.5.0b7 (Werk #5814).