Introduce the rescheduler #394
Conversation
For kubernetes-retired#118. - Currently using raw Pod as per [salt setup](https://github.com/kubernetes/kubernetes/blob/master/cluster/salt base/salt/rescheduler/rescheduler.manifest0 - Logging routed via standard streams - gcr image used, not sure if hyperkube has the rescheduler
Codecov Report
@@ Coverage Diff @@
## master #394 +/- ##
=======================================
Coverage 38.75% 38.75%
=======================================
Files 29 29
Lines 2276 2276
=======================================
Hits 882 882
Misses 1275 1275
Partials 119 119Continue to review full report at Codecov.
|
|
Thanks for the PR @c-knowles! |
cknowles
force-pushed
the
feature/rescheduler
branch
from
1dce572
to
0298a29
Compare
| containers: | ||
| - name: rescheduler | ||
| image: gcr.io/google-containers/rescheduler:v0.2.2 | ||
| # TODO: Make resource requirements depend on the size of the cluster |
There was a problem hiding this comment.
So, do we need to manually modify /etc/kubernetes/manifests/kube-rescheduler.yaml to update the resources.requests after each time worker nodes are scaled out considerably?
Mind documenting it?
There was a problem hiding this comment.
@mumoshu I copied this from kubernetes/kubernetes@c304fa1, it's what the salt setup has in the main repo. I'd be happy to just remove it perhaps as none of the others have it.
There was a problem hiding this comment.
I see; let it remain here so that we can at least notice users about it
There was a problem hiding this comment.
Note: As this pod is "burstable" in cpu/memory(requests set, but no limits set), as long as your cluster has enough capacity, rescheduler would keep working.
cknowles
force-pushed
the
feature/rescheduler
branch
from
f2bab96
to
8cdb645
Compare
|
New commits are not showing, perhaps due to org move to Kubernetes Incubator. Going to close and open a new PR. |
|
Now in #441. |
|
OK. Thanks for your effort 👍 |
For #118.