Werk #9501: Fix failing cluster start after reboot (1.5 regression)
| Component | Cluster |
| Title | Fix failing cluster start after reboot (1.5 regression) |
| Date | Oct 21, 2022 |
| Level | Trivial Change |
| Class | Bug Fix |
| Compatibility | Compatible - no manual interaction needed |
| Appliance Version | 1.5.5 |
In case the agent of a clustered appliance is monitored by a site hosted on another Checkmk server and the server queries the agent between system boot and before the cluster was initialized, then it could happend that the /omd volume was monitored accidentally before the cluster started. This was preventing the cluster startup.
In this specific case the issue was triggered by mk-job monitoring of a job executed as site user on the Checkmk appliance. To prevent this issue, we have also changed the behavior of the Checkmk 2.1 agent (see #14650).
However, the issue could also be triggered by other actions, e.g. SSH logins to a site user while the cluster is not initialized yet. To prevent this we have now explicitly disabled this early automatic mount.