Change default arp cache size on nodes #13034
Conversation
In OCP clusters with large numbers of routes (greater than the value of net.ipv4.neigh.default.gc_thresh3, which is 1024 by default) the ARP cache is not large enough to accommodate for all the entries needed by the nodes running the router pods. This change increases the cache size. bug 1425388 https://bugzilla.redhat.com/show_bug.cgi?id=1425388 Signed-off-by: Phil Cameron <[email protected]>
|
@jeremyeder is this what we want? |
|
Yep -- no known downside; double-confirmed with kernel networking folks as well. Link to mail thread is in the BZ. Hitting this starts happening near 900+ routes. Several customers have encountered it and implemented the documented tuning. It's ready to be promoted to default for OpenShift. I hope this is possible in 3.5. |
|
@eparis are you ready to do the merge? (I can't) |
|
[merge] This PR will only land in 3.6. I'm super scared to make changes in 3.5.... |
|
[Test]ing while waiting on the merge queue |
|
Evaluated for origin test up to ba84207 |
|
@eparis ack, 3.6 only -- thanks! |
|
continuous-integration/openshift-jenkins/test SUCCESS (https://ci.openshift.redhat.com/jenkins/job/test_pull_requests_origin_future/403/) (Base Commit: 2d20080) |
|
[merge] |
|
Evaluated for origin merge up to ba84207 |
|
LGTM! |
|
continuous-integration/openshift-jenkins/merge FAILURE (https://ci.openshift.redhat.com/jenkins/job/test_pull_requests_origin_future/468/) (Base Commit: 80b2f6d) |
In OCP clusters with large numbers of routes (greater than the value of
net.ipv4.neigh.default.gc_thresh3, which is 1024 by default) the ARP
cache is not large enough to accommodate for all the entries needed by
the nodes running the router pods.
This change increases the cache size.
bug 1425388
https://bugzilla.redhat.com/show_bug.cgi?id=1425388
Signed-off-by: Phil Cameron [email protected]