E2E using gce first pass#31
Conversation
|
@ravisantoshgudimetla thanks ravi, will look into it soon. |
c827f92 to
897f314
Compare
0dbc5e8 to
3f549cf
Compare
|
ok to test |
1 similar comment
|
ok to test |
|
Can one of the admins verify this patch? |
7d28eaa to
27cc1cd
Compare
|
ok to test |
|
retest this please |
|
/test all |
cb39760 to
9a2e9ac
Compare
efff065 to
42d71f7
Compare
|
/test e2e |
|
@ravisantoshgudimetla: you can't request testing unless you are a kubernetes-incubator member. DetailsIn response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
|
@ravisantoshgudimetla this PR updates vendor dir, but where are changes to glide.lock file? |
42d71f7 to
b599802
Compare
b599802 to
616d4f8
Compare
|
@aveshagarwal I removed the dependency on those vendored files. I copied the function which I need to the test so it should run independently. |
|
/test e2e |
1 similar comment
|
/test e2e |
| gcloud compute copy-files $E2E_GCE_HOME/kubeadm_preinstall.sh descheduler-$master_uuid:/tmp --zone=us-east1-b | ||
| gcloud compute copy-files $E2E_GCE_HOME/kubeadm_install.sh descheduler-$master_uuid:/tmp --zone=us-east1-b | ||
| gcloud compute copy-files $E2E_GCE_HOME/kubeadm_preinstall.sh descheduler-$node1_uuid:/tmp --zone=us-east1-b | ||
| gcloud compute copy-files $E2E_GCE_HOME/kubeadm_preinstall.sh descheduler-$node2_uuid:/tmp --zone=us-east1-b |
There was a problem hiding this comment.
WARNING: gcloud compute copy-files is deprecated. Please use gcloud compute scp instead. Note that gcloud compute scp does not have recursive copy on by default. To turn on recursion, use the --recurse flag.
There was a problem hiding this comment.
Thanks for catching it. These were not deprecated when I wrote them back in October. I can add a TODO or update them.
|
/test e2e |
616d4f8 to
e0d3b26
Compare
|
/test e2e |
1 similar comment
|
/test e2e |
|
CI job merged. |
|
@ingvagabund @ravisantoshgudimetla is it ready for merge? |
|
@ingvagabund can you give a LGTM so that @aveshagarwal can merge? |
| @@ -0,0 +1,9 @@ | |||
| #!/usr/bin/env bash | |||
There was a problem hiding this comment.
NIT: Currently the file is not needed. Installation of the gcloud is part of the JJ [1].
| @@ -0,0 +1,8 @@ | |||
| #!/usr/bin/env bash | |||
There was a problem hiding this comment.
NIT: The script is not invoked anywhere. Maybe make it part of some documentation that describes how to setup and run the descheduler e2e tests?
There was a problem hiding this comment.
This is needed if someone wants to run e2e using GCE from local system. This is similar to kube-up.sh in upstream.
| if err != nil { | ||
| t.Errorf("Error listing node with %v", err) | ||
| } | ||
| leastLoadedNode := nodeList.Items[2] |
There was a problem hiding this comment.
Nit: We should check if the cluster has at least 3 nodes. Otherwise this will panic.
Btw. how do you know the third node in the list is actually the least loaded node?
There was a problem hiding this comment.
Nit: We should check if the cluster has at least 3 nodes. Otherwise this will panic.
We would have failed in gcloud_cluster_create.sh if we don't have 3 nodes. I thought failing early is better.
Btw. how do you the third node in the list is actually the least loaded node?
I have added a comment here.
There was a problem hiding this comment.
Ok, let's put the comment here. Smth like:
// Assumption: kubeadm brings all the master components onto master node. So, the last node would have least utilization.
e0d3b26 to
16a452e
Compare
16a452e to
de76f9b
Compare
|
/lgtm |
…ream bug 1836971: Sync with upstream
…ith_travis_gcloud E2E using gce first pass
@aveshagarwal I am able to create a 3 node cluster in gce with pod being evicted for lownodeutilization case. Please review when you find time. I believe, travis ci might complain because of encrypted file(client.json.enc) which will probably not be decrypted on this repo.
Some changes:
make test.make test, we need to ensure that gcloud_create_cluster.sh has been run. I have included that in the run-tests.sh(file renamed from run-unittests.sh)Sample Output after running e2e:
evicting pods from node %#v with usage: %#v descheduler-b148f4c6-bd94-11e7-8e84-c85b76333877 map[cpu:2264 memory:263.78542244169597 pods:81.81818181818181]