Details
-
Wish
-
Status: Open
-
Minor
-
Resolution: Unresolved
-
None
-
None
-
None
Description
Kernel 4.19 release adds a new controller that attempts to guarantee minimum I/O latency targets for cgroups.
As long as everybody is meeting their latency target the controller doesn't do anything, but once a group starts missing its target it will attempt to maintain average IO latencies below the configured latency target, throttling anybody with a higher latency target than the victimized group. For more details see the documentation.
This would be useful, for example, to protect etcd workloads running in a Mesos UCR container. Please refer to etcd Disk tuning section in this link for more information.