[OCPCLOUD-1385] Ensure removal of placement-groups during cluster destroy on AWS #5528
Conversation
pkg/destroy/aws/aws.go
Outdated
| }) | ||
| if err != nil { | ||
| if err != nil { | ||
| if err.(awserr.Error).Code() == "InvalidPlacementGroup.Unknown" { |
There was a problem hiding this comment.
This magic string came from a manual test, I tried to delete a placement group which I had already deleted and got this response body, so I think I have the correct string here based on this example
<?xml version="1.0" encoding="UTF-8"?>
<Response><Errors><Error><Code>InvalidPlacementGroup.Unknown</Code><Message>The Placement Group 'pg-04e435e5cedc43563' is unknown.</Message></Error></Errors><RequestID>cf6e03f4-35fe-40cc-bd20-15bd76759d40</RequestID></Response>
pkg/destroy/aws/aws.go
Outdated
| GroupIds: []*string{aws.String(id)}, | ||
| }) | ||
| if err != nil { | ||
| if err != nil { |
There was a problem hiding this comment.
This is a duplicate check for err != nil.
There was a problem hiding this comment.
Good catch, have resolved and pushed over the top
pkg/destroy/aws/aws.go
Outdated
| }) | ||
| if err != nil { | ||
| if err != nil { | ||
| if err.(awserr.Error).Code() == "InvalidPlacementGroup.Unknown" { |
There was a problem hiding this comment.
It is terribly unfortunate that we have to use these magic string values to determine whether a resource has been deleted. It would be so much nicer if the AWS SDK exposed in its public interface the HTTP status code. There is nothing to do here, it is just a gripe.
There was a problem hiding this comment.
My comment may not have loaded while you reviewed, but I made a note as well about where this came from, I think I got the correct magic string #5528 (comment)
There was a problem hiding this comment.
Yep. I also verified that it was correct by trying to delete a placement group that I had already deleted.
JoelSpeed
force-pushed
the
remove-placement-groups
branch
from
5b4fdba
to
7e746d6
Compare
JoelSpeed
force-pushed
the
remove-placement-groups
branch
from
7e746d6
to
2785e7d
Compare
openshift-ci
bot
added
the
do-not-merge/hold
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: staebler The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
openshift-ci
bot
added
the
approved
|
/uncc |
|
/hold cancel |
openshift-ci
bot
removed
the
do-not-merge/hold
|
/retest-required Please review the full test history for this PR and help us cut down flakes. |
4 similar comments
|
/retest-required Please review the full test history for this PR and help us cut down flakes. |
|
/retest-required Please review the full test history for this PR and help us cut down flakes. |
|
/retest-required Please review the full test history for this PR and help us cut down flakes. |
|
/retest-required Please review the full test history for this PR and help us cut down flakes. |
|
@JoelSpeed: The following tests failed, say
Full PR test history. Your PR dashboard. 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. I understand the commands that are listed here. |
|
/retest-required Please review the full test history for this PR and help us cut down flakes. |
2 similar comments
|
/retest-required Please review the full test history for this PR and help us cut down flakes. |
|
/retest-required Please review the full test history for this PR and help us cut down flakes. |
As a part of the effort to introduce AWS EFA into openshift, we need to also include placement groups (openshift/enhancements#995).
As part of this, Machine API will create placement groups for users if they do not exist. During cluster teardown, these currently block the destroy operation as the installer does not know how to handle the placement groups.
This PR teaches the destroy logic how to handle placement groups, with an implementation based on other resources.
I have manually tested this by creating a cluster, adding two placement groups (created by MAPI) and then destroying the cluster. It no longer blocks indefinitely and successfully removes the placement groups and completes the cluster destroy operation.