Details
-
New Feature
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Fixed
-
23.02.3
-
None
-
2023-6
Description
MCS supports cgroup v1 only with no proper checks for values that cgroups v1 returns,e.g for memory limits. There were some combinations of kernel versions/container runtimes that has unreasonable values for max memory available for containers. The fact the values changes complicates the erroneous values detection.
This patch for the issue adds cgroups v2 support and also applies some sanity check for values reported by cgroups v1 hierarchy.
4QA cgroups hardware limits detection is disabled by default so one has to set SystemConfig.CGroup, e.g.
<SystemConfig>
|
<CGroup>./</CGroup>
|
</SystemConfig>
|
This sets CGroup to the default one unless one sets a named CGroup. Both should be tested though.