Fix the batched describe insance method for instance not found case#1260
Conversation
k8s-ci-robot
added
do-not-merge/release-note-label-needed
k8s-ci-robot
added
the
needs-triage
|
This issue is currently awaiting triage. If cloud-provider-aws contributors determine this is a relevant issue, they will accept it by applying the The DetailsInstructions 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-sigs/prow repository. |
k8s-ci-robot
added
the
size/L
|
/retest |
|
@kmala: The following tests failed, say
Full PR test history. Your PR dashboard. Please help us cut down on flakes by linking to an open issue when you hit one in your PR. DetailsInstructions 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-sigs/prow repository. I understand the commands that are listed here. |
|
/lgtm |
|
@shiv-amz: changing LGTM is restricted to collaborators 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-sigs/prow repository. |
| results := make([]batcher.Result[ec2types.Instance], len(inputs)) | ||
| firstInput := inputs[0] | ||
|
|
||
| firstInput := *inputs[0] |
There was a problem hiding this comment.
why did we translate this from a pointer? can potentially cause another panic?
There was a problem hiding this comment.
if we don't translate we are overriding it in the below for loop as firstInput would just be a pointer to the first item item in slice. inputs should never have nil items for the panic to happen and there should always be atleast 1 item in the inputs.
There was a problem hiding this comment.
Should we still have a len check on inputs before this line to catch the case where inputs is empty? Unless we have a hard guarantee somewhere that inputs is not empty?
There was a problem hiding this comment.
yes, we are guaranteed to not have empty list as this gets called only when we make a describe instance call and there it verifies the instance count.
| return nil, fmt.Errorf("expected to receive a single instance only, found %d", len(input.InstanceIds)) | ||
| } | ||
| result := b.batcher.Add(ctx, input) | ||
| if result.Output == nil { |
There was a problem hiding this comment.
is if result.Err != nil a better check here? i wonder whether we're actually getting an error e.g. 404 not found wen describing a non-existent instance, it's trickier when e.g. we batch two describe instance IDs where one exists and the other is non-existent, are we getting an error in this case?
|
I think it's more complex than a nil check. Bundling the describe calls via introducing a batcher did lower down the API counts but it also adds some complexity regarding error handling. i just did a local experiment:
@sarthakkothari @shiv-amz can you confirm the batcher behavior and add related test cases to that? |
|
/lgtm Discussed with @kmala offline, merge this in as this PR will fix the gap for error handling. |
k8s-ci-robot
added
the
lgtm
|
/approve |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: yue9944882 The full list of commands accepted by this bot can be found here. The pull request process is described here DetailsNeeds approval from an approver in each of these files:
Approvers can indicate their approval by writing |
k8s-ci-robot
added
the
approved
|
/release-note-none |
k8s-ci-robot
added
release-note-none
…stream-release-1.34 Automated cherry pick of #1260: Fix the batched describe insance method for instance not
6 participants
What type of PR is this?
/kind bug
What this PR does / why we need it:
Fix the describe instance batcher when the instance is not found or the batch call return error. If the instance is not found we shouldn't be returning an array of empty object but should be returning a nil. This is causing CCM to panic when the nodes deletion in apiserver is blocked because of finalizers but the instance in the aws is deleted. This bug is introduced in the #1152
Which issue(s) this PR fixes:
Fixes #
Special notes for your reviewer:
Does this PR introduce a user-facing change?: