diff options
| author | Jiri Mencak <jmencak@redhat.com> | 2017-10-03 19:57:01 +0200 |
|---|---|---|
| committer | Jiri Mencak <jmencak@redhat.com> | 2017-10-10 13:47:35 +0200 |
| commit | 17e38ad8c62ac8f1c8f983e63181f69879a45d85 (patch) | |
| tree | a3038e8f2fe7d8d0ea7f20b77a7ea7c3841ea14e /roles/tuned/templates/openshift-control-plane | |
| parent | 2fb41eea9aa4b328d7da1c5e4b7de2e717549ed7 (diff) | |
| download | openshift-17e38ad8c62ac8f1c8f983e63181f69879a45d85.tar.gz openshift-17e38ad8c62ac8f1c8f983e63181f69879a45d85.tar.bz2 openshift-17e38ad8c62ac8f1c8f983e63181f69879a45d85.tar.xz openshift-17e38ad8c62ac8f1c8f983e63181f69879a45d85.zip | |
Separate tuned daemon setup into a role.
Currently, profiles for the tuned daemon are set only for
OpenShift node(s). This excludes the OpenShift loadbalancer.
As a result, ARP cache limits on loadbalancers are not raised.
This causes problems with HA setups where loadbalancers serve
1k+ OpenShift nodes.
This commit ensures the openshift-control-plane role is applied
to loadbalancers, masters and OpenShift infra nodes. Regular
OpenShift worker nodes get the openshift-node profile.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1498213
Diffstat (limited to 'roles/tuned/templates/openshift-control-plane')
| -rw-r--r-- | roles/tuned/templates/openshift-control-plane/tuned.conf | 25 |
1 files changed, 25 insertions, 0 deletions
diff --git a/roles/tuned/templates/openshift-control-plane/tuned.conf b/roles/tuned/templates/openshift-control-plane/tuned.conf new file mode 100644 index 000000000..f22f21065 --- /dev/null +++ b/roles/tuned/templates/openshift-control-plane/tuned.conf @@ -0,0 +1,25 @@ +# +# tuned configuration +# + +[main] +summary=Optimize systems running OpenShift control plane +include=openshift + +[sysctl] +# ktune sysctl settings, maximizing i/o throughput +# +# Minimal preemption granularity for CPU-bound tasks: +# (default: 1 msec# (1 + ilog(ncpus)), units: nanoseconds) +kernel.sched_min_granularity_ns=10000000 + +# The total time the scheduler will consider a migrated process +# "cache hot" and thus less likely to be re-migrated +# (system default is 500000, i.e. 0.5 ms) +kernel.sched_migration_cost_ns=5000000 + +# SCHED_OTHER wake-up granularity. +# +# Preemption granularity when tasks wake up. Lower the value to improve +# wake-up latency and throughput for latency critical tasks. +kernel.sched_wakeup_granularity_ns = 4000000 |