Werk #22645: Value store: services crashed on every run after a timeout
| Component | Core & setup | ||||||
| Title | Value store: services crashed on every run after a timeout | ||||||
| Date | Oct 26, 2026 | ||||||
| Level | Trivial Change | ||||||
| Class | Bug Fix | ||||||
| Compatibility | Compatible - no manual interaction needed | ||||||
| Checkmk versions & editions |
|
With the Checkmk Micro Core, if the Checkmk service of a host timed out
while the counter values of its services were being saved, one of the values
could be replaced by an error message starting with <Exception in. The same
could happen with other errors raised at that point, also with the Nagios
core.
From then on, the affected service crashed on every check execution with
SyntaxError: invalid syntax. This was e.g. seen for the service "Disk IO
SUMMARY". The crash persisted across restarts and updates of the site.
Checkmk now aborts saving the counter values if the timeout occurs at that point, and the Checkmk service reports the timeout. For any other error, the previous value of the affected counter is kept and a warning is logged.
Counter values that are already affected are discarded with a warning the next time they are read. The service then initializes the counter again and recovers by itself. You do not need to delete any counter files.
With the Checkmk Micro Core, these warnings are written to var/log/cmc.log.
With the Nagios core, they appear in the output of the Checkmk service.