fix: avoid recreate vmss cache in race condition

[andyzhangx]

What type of PR is this?

/kind bug

What this PR does / why we need it:

fix: avoid recreate vmss cache in race condition
In race condition, there could be duplicated getVMSSVMCache calls, and it would end up with creating multiple vmss cache objects, and finally cause VMSS list storm, related logs:

I1018 03:18:45.242671       1 azure_vmss.go:204] Couldn't find VMSS VM with nodeName aks-agentpool-34156162-vmss00000j, refreshing the cache(vmss: aks-agentpool-34156162-vmss, rg: aksdiskscaleai2f68d-nodegroup)
I1018 03:18:45.242671       1 azure_vmss.go:204] Couldn't find VMSS VM with nodeName aks-agentpool-34156162-vmss00000j, refreshing the cache(vmss: aks-agentpool-34156162-vmss, rg: aksdiskscaleai2f68d-nodegroup)

I1018 03:18:47.272599       1 azure_vmss.go:204] Couldn't find VMSS VM with nodeName aks-agentpool-34156162-vmss00000x, refreshing the cache(vmss: aks-agentpool-34156162-vmss, rg: aksdiskscaleai2f68d-nodegroup)
I1018 03:18:47.272599       1 azure_vmss.go:204] Couldn't find VMSS VM with nodeName aks-agentpool-34156162-vmss00000x, refreshing the cache(vmss: aks-agentpool-34156162-vmss, rg: aksdiskscaleai2f68d-nodegroup)

I1018 03:18:51.474445       1 azure_vmss.go:204] Couldn't find VMSS VM with nodeName aks-agentpool-34156162-vmss00000l, refreshing the cache(vmss: aks-agentpool-34156162-vmss, rg: aksdiskscaleai2f68d-nodegroup)
I1018 03:18:51.474445       1 azure_vmss.go:204] Couldn't find VMSS VM with nodeName aks-agentpool-34156162-vmss00000l, refreshing the cache(vmss: aks-agentpool-34156162-vmss, rg: aksdiskscaleai2f68d-nodegroup)

Which issue(s) this PR fixes:

Fixes #

Special notes for your reviewer:

Does this PR introduce a user-facing change?

fix: avoid recreate vmss cache in race condition

Additional documentation e.g., KEPs (Kubernetes Enhancement Proposals), usage docs, etc.:

fix: avoid recreate vmss cache in race condition

[netlify]

✅ Deploy Preview for kubernetes-sigs-cloud-provide-azure canceled.

Name	Link
🔨 Latest commit	fd6696b
🔍 Latest deploy log	https://app.netlify.com/sites/kubernetes-sigs-cloud-provide-azure/deploys/6354a7caba23d9000970699a

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: andyzhangx

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:

• OWNERS [andyzhangx]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[edreed]

Have you considered using singleflight to suppress duplicate calls rather than adding a new lock?

[andyzhangx]

@edreed using singleflight requires refactor of vmss cache, adding a new lock would be quite small and safe change, let's roll out this fix first

[edreed]

It shouldn't require a refactor for use in this context, e.g.

v, err, _ := ss.group.Do(cacheKey, func () (interface{}, error) {
	cache, err := ss.newVMSSVirtualMachinesCache(resourceGroup, vmssName, cacheKey)
	if err != nil {
		return nil, err
	}
	ss.vmssVMCache.Store(cacheKey, cache)
	return cache, nil
});
if err != nil {
	return "", nil, err
}
cache := v.(*azcache.TimedCache)
return cacheKey, cache, nil

where ss.group is a singleflight.Group.

[andyzhangx]

thanks, using singleflight.Group now in the PR, I think this PR could fix the vmss list storm since it does not have lock in the before, lots of new vmss cache are created simultaneously which cause the vmss list storm.

[coveralls]

Coverage decreased (-0.006%) to 79.855% when pulling fd6696b on andyzhangx:avoid-recreate-vmss-cache into c60daf9 on kubernetes-sigs:master.

[MartinForReal]

/test pull-cloud-provider-azure-e2e-ccm-capz

[MartinForReal]

/lgtm

[MartinForReal]

/cherrypick release-1.25

[k8s-infra-cherrypick-robot]

@MartinForReal: once the present PR merges, I will cherry-pick it on top of release-1.25 in a new PR and assign it to you.

In response to this:

/cherrypick release-1.25

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.

[MartinForReal]

/cherrypick release-1.24

[k8s-infra-cherrypick-robot]

@MartinForReal: once the present PR merges, I will cherry-pick it on top of release-1.24 in a new PR and assign it to you.

In response to this:

/cherrypick release-1.24

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.

[MartinForReal]

/cherrypick release-1.23

[k8s-infra-cherrypick-robot]

@MartinForReal: once the present PR merges, I will cherry-pick it on top of release-1.23 in a new PR and assign it to you.

In response to this:

/cherrypick release-1.23

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.

[MartinForReal]

/cherrypick release-1.1

[k8s-infra-cherrypick-robot]

@MartinForReal: once the present PR merges, I will cherry-pick it on top of release-1.1 in a new PR and assign it to you.

In response to this:

/cherrypick release-1.1

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.

[k8s-infra-cherrypick-robot]

@MartinForReal: #2589 failed to apply on top of branch "release-1.25":

Applying: fix: avoid recreate vmss cache in race condition
Applying: feat: use singleflight.Group instead of lock to fix race condition
Using index info to reconstruct a base tree...
M	go.mod
M	pkg/provider/azure_vmss.go
M	pkg/provider/azure_vmss_cache.go
Falling back to patching base and 3-way merge...
Auto-merging pkg/provider/azure_vmss_cache.go
Auto-merging pkg/provider/azure_vmss.go
Auto-merging go.mod
Applying: doc: fix spelling error
Using index info to reconstruct a base tree...
M	pkg/provider/azure_controller_common.go
Falling back to patching base and 3-way merge...
Auto-merging pkg/provider/azure_controller_common.go
CONFLICT (content): Merge conflict in pkg/provider/azure_controller_common.go
error: Failed to merge in the changes.
hint: Use 'git am --show-current-patch=diff' to see the failed patch
Patch failed at 0003 doc: fix spelling error
When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".

In response to this:

/cherrypick release-1.25

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.

[k8s-infra-cherrypick-robot]

@MartinForReal: #2589 failed to apply on top of branch "release-1.24":

Applying: fix: avoid recreate vmss cache in race condition
Applying: feat: use singleflight.Group instead of lock to fix race condition
Using index info to reconstruct a base tree...
M	go.mod
M	pkg/provider/azure_vmss.go
M	pkg/provider/azure_vmss_cache.go
Falling back to patching base and 3-way merge...
Auto-merging pkg/provider/azure_vmss_cache.go
Auto-merging pkg/provider/azure_vmss.go
Auto-merging go.mod
CONFLICT (content): Merge conflict in go.mod
error: Failed to merge in the changes.
hint: Use 'git am --show-current-patch=diff' to see the failed patch
Patch failed at 0002 feat: use singleflight.Group instead of lock to fix race condition
When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".

In response to this:

/cherrypick release-1.24

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.

[k8s-infra-cherrypick-robot]

@MartinForReal: #2589 failed to apply on top of branch "release-1.23":

Applying: fix: avoid recreate vmss cache in race condition
Applying: feat: use singleflight.Group instead of lock to fix race condition
Using index info to reconstruct a base tree...
M	go.mod
M	pkg/provider/azure_vmss.go
M	pkg/provider/azure_vmss_cache.go
Falling back to patching base and 3-way merge...
Auto-merging pkg/provider/azure_vmss_cache.go
Auto-merging pkg/provider/azure_vmss.go
Auto-merging go.mod
Applying: doc: fix spelling error
Using index info to reconstruct a base tree...
M	pkg/provider/azure_controller_common.go
Falling back to patching base and 3-way merge...
Auto-merging pkg/provider/azure_controller_common.go
CONFLICT (content): Merge conflict in pkg/provider/azure_controller_common.go
error: Failed to merge in the changes.
hint: Use 'git am --show-current-patch=diff' to see the failed patch
Patch failed at 0003 doc: fix spelling error
When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".

In response to this:

/cherrypick release-1.23

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.

[k8s-infra-cherrypick-robot]

@MartinForReal: #2589 failed to apply on top of branch "release-1.1":

Applying: fix: avoid recreate vmss cache in race condition
Applying: feat: use singleflight.Group instead of lock to fix race condition
Using index info to reconstruct a base tree...
M	go.mod
M	pkg/provider/azure_vmss.go
M	pkg/provider/azure_vmss_cache.go
Falling back to patching base and 3-way merge...
Auto-merging pkg/provider/azure_vmss_cache.go
Auto-merging pkg/provider/azure_vmss.go
CONFLICT (content): Merge conflict in pkg/provider/azure_vmss.go
Auto-merging go.mod
CONFLICT (content): Merge conflict in go.mod
error: Failed to merge in the changes.
hint: Use 'git am --show-current-patch=diff' to see the failed patch
Patch failed at 0002 feat: use singleflight.Group instead of lock to fix race condition
When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".

In response to this:

/cherrypick release-1.1

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.