Run calico-manifest-deployment as a periodic job#604
Conversation
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: kishen-v The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
ppc64le-cloud-bot
added
the
size/M
c9f936c to
ed1d4ad
Compare
mkumatag
left a comment
mkumatag
left a comment
There was a problem hiding this comment.
how expensive this will be? can we run in the upstream prow itself?
|
Just confirmed from the estimator, it roughly costs $0.15 an hour (no workers deployed as we'd just want to see if the node and the controller pods deploy successfully) , so even if the test runs for 2 hours at the most, we'd be looking at $0.60 a week, and ~$2.40-$3 a month, or the best case scenario is that it completes within 30 minutes, which should cost us $1.2-$1.5 a month. The estimate is for a 0.25C x 8GB deployment. I'll quickly make changes to set the /hold |
ppc64le-cloud-bot
added
the
do-not-merge/hold
nice, let us use upstream prow for this then, slowly we can reduce the content running on this and move away one day completely. |
|
/unhold |
ppc64le-cloud-bot
removed
the
do-not-merge/hold
| --build-version $K8S_BUILD_VERSION --release-marker=$K8S_BUILD_VERSION \ | ||
| --workers-count 0 --auto-approve --cluster-name calico-master-manifest-$TIMESTAMP --playbook install-k8s.yml --up --down --retry-on-tf-failure 3 \ | ||
| --break-kubetest-on-upfail true --ignore-destroy-errors \ | ||
| --extra-vars=cni_provider:"calico-master" --extra-vars=directory:"release" --test=exec -- \ |
There was a problem hiding this comment.
@kishen-v Is passing cni_provider:"calico-master" to avoid Install networking task ininstall-k8s.yml playbook?
There was a problem hiding this comment.
You're right.. The CNI installation is something we'd wanted to handle out of scope as the tests surround the same..
I feel cni_provider:"none" will be more appropriate.
if this is not a resource heavy - can we move to upstream k8s prow? |
4 participants
/cc @Rajalakshmi-Girish.
Notes: Do we need to run the job less frequently? Maybe once a week?