Bug #5619
closed
Process checking for cf-agent on windows suicide itself
Added by Nicolas CHARLES about 10 years ago.
Updated almost 10 years ago.
Category:
System techniques
Description
In the system technique, we check the cfengine processes, and kill them if they are too numerous and restart them
However, on windows we are not restarting properly the services (with command on cf-execd rather than service management)
Issue is that the process table is cached, so the restart-class is never defined.
On Linux, this is minigated by cron, so there is no issue about it. On Windows, there is no cron, so the executor daemon is never ever restarted.
It is also worth noting that on CFEngine masterfiles, the check on binaries is made only not on windows, as executor and monitor are services, so there can be more than one running
So the most obvious solution here is to remove altogether this check on windows for execd, and controlling only the agent execution
- Status changed from In progress to Pending technical review
- Assignee changed from Nicolas CHARLES to Jonathan CLARKE
- Pull Request set to https://github.com/Normation/rudder-techniques/pull/533
- Target version changed from 2.10.7 to 2.10.8
- Status changed from Pending technical review to Pending release
- % Done changed from 0 to 100
Applied in changeset policy-templates:commit:4cbc09cfc86f55d27f3655eb1488fe973d062512.
Applied in changeset policy-templates:commit:f9bac2da7c6e1bffd2fbcebd05f69d1ecfa1fcb4.
- Status changed from Pending release to Released
This bug has been fixed in Rudder 2.10.8 and 2.11.5, which were released today (16/12/14)
Also available in: Atom
PDF
Updated by 
Updated by 
Updated by Jonathan CLARKE