-
Notifications
You must be signed in to change notification settings - Fork 39.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Version bump to etcd v3.2.13, grpc v1.7.5 #57480
Conversation
@shyamjvs Is there any to kick off a scalability test on this PR? We'd like to ensure it's passing before merging if possible since the last grpc bump caused scalability test instability on master and we'd prefer ensure we've fixed that issue before merging. |
@jpbetz etcd part LGTM. I'd also like to know how kubernetes encountered #51099. If Thanks. |
/retest |
/test pull-kubernetes-kubemark-e2e-gce-big |
/test pull-kubernetes-bazel-test |
Both big kubemark tests passed, but the suite timedout. My feeling is that it's long build and we can ignore it. @porridge ? |
Yes, the timeout of |
On second thought, it's actually somewhat suspicious that 80 minutes were not enough for a test which passes in ~50 on master. I'll take a closer look. |
In master we are not building - with presubmit we are doing building. It used to take 30m+ in the past, maybe it's still the case. |
Right, just compared and the additional ~30 minutes are from quick-release. Let me fix the timeouts. |
/test pull-kubernetes-kubemark-e2e-gce-big |
FWIW, those pull-kubernetes-bazel-test do not look like flakes. |
Clean rebase. Adding back lgtm. |
/test all [submit-queue is verifying that this PR is safe to merge] |
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions here. |
FTR:
|
Thanks for following up @porridge |
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>. Update etcd server version to 3.2.14 This upgrades the default etcd version used by kubernetes to 3.2.14 We previously [bumped the etcd client to 3.2.14](#57480). Fixes #56438 ```release-note Upgrade default etcd server version to 3.2.14 ``` cc @gyuho
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>. Update etcd server version to 3.2.14 This upgrades the default etcd version used by kubernetes to 3.2.14 We previously [bumped the etcd client to 3.2.14](kubernetes/kubernetes#57480). Fixes kubernetes/kubernetes#56438 ```release-note Upgrade default etcd server version to 3.2.14 ``` cc @gyuho Kubernetes-commit: 0f6354e81b16030f7c2dd9c65a29cd1f5b5e43b2
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>. Update etcd server version to 3.2.14 This upgrades the default etcd version used by kubernetes to 3.2.14 We previously [bumped the etcd client to 3.2.14](kubernetes/kubernetes#57480). Fixes kubernetes/kubernetes#56438 ```release-note Upgrade default etcd server version to 3.2.14 ``` cc @gyuho Kubernetes-commit: 0f6354e81b16030f7c2dd9c65a29cd1f5b5e43b2
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>. Update kubeadm supported etcd version to 3.2.14 in 1.10 **What this PR does / why we need it**: Kubernetes will upgrade to etcd server 3.2.14 in 1.10 cycle (#58645), update DefaultEtcdVersion in kubeadm to this version. **Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*: relevant PR: #57480 #58645 fixes: kubernetes/kubeadm#621 **Special notes for your reviewer**: /cc @kubernetes/sig-cluster-lifecycle-pr-reviews **Release note**: ```release-note NONE ``` kubeadm don't need to advertise this in release notes.
@jpbetz have you figured out what was causing the "context cancelled" errors? When trying to pickup the new etcd (3.2.13 or 3.2.16), i'm seeing this panic in k8s tests: https://gist.github.com/mfojtik/a5109e6c752f2569b99b5dc90e5d1801 EDIT: The panic is timeout as in openshift/origin we set the default unit test timeout to 120s. However this test (pkg/master/master_test.go:
|
FYI - today while looking at apiserver mem usage in our 5k-node scalability tests, I noticed a huge drop across runs 93 and 95. From the diff, this change seems to have most likely caused the improvement :) Do we know why we see such improvement? What changed? |
@shyamjvs I suspect it is a combination of things. In the grpc v1.3.0 -> v1.7.5 upgrade there are various performance improvements (see https://github.com/grpc/grpc-go/releases/tag/v1.5.0, https://github.com/grpc/grpc-go/releases/tag/v1.4.0) and there are quite a few etcd 3.1-3.2 client improvements (@gyuho any that you know of that would improve memory utilization this much?). @cheftako It would be great to be able to run the memory analysis tool you've been trying out on scalability runs like this. Ideally, a tool like that would tell us exactly what changed from baseline. |
We did some performance work for 3.3 (https://github.com/coreos/etcd/blob/master/CHANGELOG-3.3.md#improved-1), but doubt client upgrade would have much effect. Yeah, I would be interested in what have caused this :) |
is this PR backportable? |
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>. Update etcd server version to 3.2.14 This upgrades the default etcd version used by kubernetes to 3.2.14 We previously [bumped the etcd client to 3.2.14](kubernetes/kubernetes#57480). Fixes kubernetes/kubernetes#56438 ```release-note Upgrade default etcd server version to 3.2.14 ``` cc @gyuho Kubernetes-commit: 0f6354e81b16030f7c2dd9c65a29cd1f5b5e43b2
Reapply #57160 but with etcd 3.2.13, which includes etcd-io/etcd#9047 to fix #51099.
We need to scalability test this PR before merging it since the previous attempt to version bump to grpc v1.7+ resulted in a scalability test failure after the PR was merged to master, and we don't want to repeat that. No, no we don't.
Thanks @gyuho for fixing the etcd grpc issue and releasing etcd-3.2.13 on short notice.
Release note: