fix: deadlock in controller #14304

agaudreault · 2023-07-01T02:39:21Z

When the config resource.ignoreResourceUpdatesEnabled: 'true' is true and the application controller log level is set to debug, the controller hangs.


1 @ 0x43bf36 0x44d6af 0x44d686 0x46c4e6 0x1eb05f9 0x1eb05d4 0x21f8229 0x1eafe74 0x1eafd08 0x1ea9f85 0x1a9b137 0x131c59b 0x131c677 0x131d3f8 0x131d369 0x131d29c 0x1a9af32 0x1ea97ff 0x470861
#	0x46c4e5	sync.runtime_SemacquireRWMutexR+0x25								/usr/local/go/src/runtime/sema.go:82
#	0x1eb05f8	sync.(*RWMutex).RLock+0x78									/usr/local/go/src/sync/rwmutex.go:71
#	0x1eb05d3	github.com/argoproj/gitops-engine/pkg/cache.(*clusterCache).GetClusterInfo+0x53			/go/pkg/mod/github.com/argoproj/gitops-engine@v0.7.1-0.20230607163028-425d65e07695/pkg/cache/cluster.go:1058
#	0x21f8228	github.com/argoproj/argo-cd/v2/controller/cache.(*liveStateCache).getCluster.func2+0x4a8	/go/src/github.com/argoproj/argo-cd/controller/cache/cache.go:516
#	0x1eafe73	github.com/argoproj/gitops-engine/pkg/cache.(*clusterCache).onNodeUpdated+0x93			/go/pkg/mod/github.com/argoproj/gitops-engine@v0.7.1-0.20230607163028-425d65e07695/pkg/cache/cluster.go:1021
#	0x1eafd07	github.com/argoproj/gitops-engine/pkg/cache.(*clusterCache).processEvent+0x3a7			/go/pkg/mod/github.com/argoproj/gitops-engine@v0.7.1-0.20230607163028-425d65e07695/pkg/cache/cluster.go:1014
#	0x1ea9f84	github.com/argoproj/gitops-engine/pkg/cache.(*clusterCache).watchEvents.func1+0x6c4		/go/pkg/mod/github.com/argoproj/gitops-engine@v0.7.1-0.20230607163028-425d65e07695/pkg/cache/cluster.go:618
#	0x1a9b136	github.com/argoproj/gitops-engine/pkg/utils/kube.RetryUntilSucceed.func1+0xf6			/go/pkg/mod/github.com/argoproj/gitops-engine@v0.7.1-0.20230607163028-425d65e07695/pkg/utils/kube/kube.go:411
#	0x131c59a	k8s.io/apimachinery/pkg/util/wait.ConditionFunc.WithContext.func1+0x1a				/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:220
#	0x131c676	k8s.io/apimachinery/pkg/util/wait.runConditionWithCrashProtectionWithContext+0x56		/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:233
#	0x131d3f7	k8s.io/apimachinery/pkg/util/wait.poll+0x37							/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:580
#	0x131d368	k8s.io/apimachinery/pkg/util/wait.PollImmediateUntilWithContext+0x48				/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:545
#	0x131d29b	k8s.io/apimachinery/pkg/util/wait.PollImmediateUntil+0x7b					/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:536
#	0x1a9af31	github.com/argoproj/gitops-engine/pkg/utils/kube.RetryUntilSucceed+0x131			/go/pkg/mod/github.com/argoproj/gitops-engine@v0.7.1-0.20230607163028-425d65e07695/pkg/utils/kube/kube.go:409
#	0x1ea97fe	github.com/argoproj/gitops-engine/pkg/cache.(*clusterCache).watchEvents+0x2de			/go/pkg/mod/github.com/argoproj/gitops-engine@v0.7.1-0.20230607163028-425d65e07695/pkg/cache/cluster.go:534

1 @ 0x43bf36 0x44d6af 0x44d686 0x46c4e6 0x1eb05f9 0x1eb05d4 0x21fb688 0x21044ee 0x470861
#	0x46c4e5	sync.runtime_SemacquireRWMutexR+0x25							/usr/local/go/src/runtime/sema.go:82
#	0x1eb05f8	sync.(*RWMutex).RLock+0x78								/usr/local/go/src/sync/rwmutex.go:71
#	0x1eb05d3	github.com/argoproj/gitops-engine/pkg/cache.(*clusterCache).GetClusterInfo+0x53		/go/pkg/mod/github.com/argoproj/gitops-engine@v0.7.1-0.20230607163028-425d65e07695/pkg/cache/cluster.go:1058
#	0x21fb687	github.com/argoproj/argo-cd/v2/controller/cache.(*liveStateCache).GetClustersInfo+0x2c7	/go/src/github.com/argoproj/argo-cd/controller/cache/cache.go:797
#	0x21044ed	github.com/argoproj/argo-cd/v2/controller/metrics.(*clusterCollector).Run+0xcd		/go/src/github.com/argoproj/argo-cd/controller/metrics/clustercollector.go:71

1 @ 0x43bf36 0x44d6af 0x44d686 0x46c4e6 0x1eb05f9 0x1eb05d4 0x21fb688 0x2279676 0x22795cf 0x470861
#	0x46c4e5	sync.runtime_SemacquireRWMutexR+0x25							/usr/local/go/src/runtime/sema.go:82
#	0x1eb05f8	sync.(*RWMutex).RLock+0x78								/usr/local/go/src/sync/rwmutex.go:71
#	0x1eb05d3	github.com/argoproj/gitops-engine/pkg/cache.(*clusterCache).GetClusterInfo+0x53		/go/pkg/mod/github.com/argoproj/gitops-engine@v0.7.1-0.20230607163028-425d65e07695/pkg/cache/cluster.go:1058
#	0x21fb687	github.com/argoproj/argo-cd/v2/controller/cache.(*liveStateCache).GetClustersInfo+0x2c7	/go/src/github.com/argoproj/argo-cd/controller/cache/cache.go:797
#	0x2279675	github.com/argoproj/argo-cd/v2/controller.(*clusterInfoUpdater).updateClusters+0x55	/go/src/github.com/argoproj/argo-cd/controller/clusterinfoupdater.go:63
#	0x22795ce	github.com/argoproj/argo-cd/v2/controller.(*clusterInfoUpdater).Run+0xae		/go/src/github.com/argoproj/argo-cd/controller/clusterinfoupdater.go:56

1 @ 0x43bf36 0x44d6af 0x44d686 0x46c4e6 0x1ead50b 0x1ead4e6 0x21f95ec 0x2267d75 0x2266fc5 0x22730d5 0x226bc89 0x131c37e 0x131c236 0x131c129 0x131c065 0x470861
#	0x46c4e5	sync.runtime_SemacquireRWMutexR+0x25									/usr/local/go/src/runtime/sema.go:82
#	0x1ead50a	sync.(*RWMutex).RLock+0xaa										/usr/local/go/src/sync/rwmutex.go:71
#	0x1ead4e5	github.com/argoproj/gitops-engine/pkg/cache.(*clusterCache).IterateHierarchy+0x85			/go/pkg/mod/github.com/argoproj/gitops-engine@v0.7.1-0.20230607163028-425d65e07695/pkg/cache/cluster.go:855
#	0x21f95eb	github.com/argoproj/argo-cd/v2/controller/cache.(*liveStateCache).IterateHierarchy+0xcb			/go/src/github.com/argoproj/argo-cd/controller/cache/cache.go:587
#	0x2267d74	github.com/argoproj/argo-cd/v2/controller.(*ApplicationController).getResourceTree+0x7b4		/go/src/github.com/argoproj/argo-cd/controller/appcontroller.go:466
#	0x2266fc4	github.com/argoproj/argo-cd/v2/controller.(*ApplicationController).setAppManagedResources+0xa4		/go/src/github.com/argoproj/argo-cd/controller/appcontroller.go:386
#	0x22730d4	github.com/argoproj/argo-cd/v2/controller.(*ApplicationController).processAppRefreshQueueItem+0x1234	/go/src/github.com/argoproj/argo-cd/controller/appcontroller.go:1444
#	0x226bc88	github.com/argoproj/argo-cd/v2/controller.(*ApplicationController).Run.func3+0x28			/go/src/github.com/argoproj/argo-cd/controller/appcontroller.go:738
#	0x131c37d	k8s.io/apimachinery/pkg/util/wait.BackoffUntil.func1+0x3d						/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:155
#	0x131c235	k8s.io/apimachinery/pkg/util/wait.BackoffUntil+0xb5							/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:156
#	0x131c128	k8s.io/apimachinery/pkg/util/wait.JitterUntil+0x88							/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:133
#	0x131c064	k8s.io/apimachinery/pkg/util/wait.Until+0x24								/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:90

96 @ 0x43bf36 0x46c72c 0x46c70c 0x47948c 0x13222e5 0x2271ee7 0x226bc89 0x131c37e 0x131c236 0x131c129 0x131c065 0x470861
#	0x46c70b	sync.runtime_notifyListWait+0x12b									/usr/local/go/src/runtime/sema.go:527
#	0x47948b	sync.(*Cond).Wait+0x8b											/usr/local/go/src/sync/cond.go:70
#	0x13222e4	k8s.io/client-go/util/workqueue.(*Type).Get+0xa4							/go/pkg/mod/k8s.io/client-go@v0.24.2/util/workqueue/queue.go:157
#	0x2271ee6	github.com/argoproj/argo-cd/v2/controller.(*ApplicationController).processAppRefreshQueueItem+0x46	/go/src/github.com/argoproj/argo-cd/controller/appcontroller.go:1319
#	0x226bc88	github.com/argoproj/argo-cd/v2/controller.(*ApplicationController).Run.func3+0x28			/go/src/github.com/argoproj/argo-cd/controller/appcontroller.go:738
#	0x131c37d	k8s.io/apimachinery/pkg/util/wait.BackoffUntil.func1+0x3d						/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:155
#	0x131c235	k8s.io/apimachinery/pkg/util/wait.BackoffUntil+0xb5							/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:156
#	0x131c128	k8s.io/apimachinery/pkg/util/wait.JitterUntil+0x88							/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:133
#	0x131c064	k8s.io/apimachinery/pkg/util/wait.Until+0x24								/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:90

3 @ 0x43bf36 0x44d6af 0x44d686 0x46c4e6 0x1ea6385 0x1ea6359 0x21f9e13 0x227bda2 0x2284285 0x2272efe 0x226bc89 0x131c37e 0x131c236 0x131c129 0x131c065 0x470861
#	0x46c4e5	sync.runtime_SemacquireRWMutexR+0x25									/usr/local/go/src/runtime/sema.go:82
#	0x1ea6384	sync.(*RWMutex).RLock+0x64										/usr/local/go/src/sync/rwmutex.go:71
#	0x1ea6358	github.com/argoproj/gitops-engine/pkg/cache.(*clusterCache).GetAPIResources+0x38			/go/pkg/mod/github.com/argoproj/gitops-engine@v0.7.1-0.20230607163028-425d65e07695/pkg/cache/cluster.go:289
#	0x21f9e12	github.com/argoproj/argo-cd/v2/controller/cache.(*liveStateCache).GetVersionsInfo+0x72			/go/src/github.com/argoproj/argo-cd/controller/cache/cache.go:635
#	0x227bda1	github.com/argoproj/argo-cd/v2/controller.(*appStateManager).getRepoObjs+0x561				/go/src/github.com/argoproj/argo-cd/controller/state.go:148
#	0x2284284	github.com/argoproj/argo-cd/v2/controller.(*appStateManager).CompareAppState+0x5284			/go/src/github.com/argoproj/argo-cd/controller/state.go:400
#	0x2272efd	github.com/argoproj/argo-cd/v2/controller.(*ApplicationController).processAppRefreshQueueItem+0x105d	/go/src/github.com/argoproj/argo-cd/controller/appcontroller.go:1434
#	0x226bc88	github.com/argoproj/argo-cd/v2/controller.(*ApplicationController).Run.func3+0x28			/go/src/github.com/argoproj/argo-cd/controller/appcontroller.go:738
#	0x131c37d	k8s.io/apimachinery/pkg/util/wait.BackoffUntil.func1+0x3d						/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:155
#	0x131c235	k8s.io/apimachinery/pkg/util/wait.BackoffUntil+0xb5							/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:156
#	0x131c128	k8s.io/apimachinery/pkg/util/wait.JitterUntil+0x88							/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:133
#	0x131c064	k8s.io/apimachinery/pkg/util/wait.Until+0x24								/go/pkg/mod/k8s.io/apimachinery@v0.24.2/pkg/util/wait/wait.go:90

codecov · 2023-07-01T02:50:55Z

Codecov Report

Patch and project coverage have no change.

Comparison is base (8032601) 49.75% compared to head (e48bf27) 49.76%.

Additional details and impacted files

@@           Coverage Diff           @@
##           master   #14304   +/-   ##
=======================================
  Coverage   49.75%   49.76%           
=======================================
  Files         261      261           
  Lines       44659    44659           
=======================================
+ Hits        22222    22225    +3     
+ Misses      20251    20249    -2     
+ Partials     2186     2185    -1

Impacted Files	Coverage Δ
controller/cache/cache.go	`25.10% <0.00%> (ø)`

... and 1 file with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

jannfis · 2023-07-01T15:43:17Z

Please submit PR against master, we'll cherry-pick into release-2.8.

crenshaw-dev · 2023-07-01T16:04:39Z

Changed the target branch, but diff is gonna be weird until rebase.

crenshaw-dev · 2023-07-01T16:04:56Z

/cherry-pick release-2.8

Signed-off-by: Alexandre Gaudreault <alexandre.gaudreault@logmein.com>

jannfis

LGTM, thanks

Signed-off-by: Alexandre Gaudreault <alexandre.gaudreault@logmein.com>

Signed-off-by: Alexandre Gaudreault <alexandre.gaudreault@logmein.com> Co-authored-by: Alexandre Gaudreault <alexandre.gaudreault@logmein.com>

Signed-off-by: Alexandre Gaudreault <alexandre.gaudreault@logmein.com>

crenshaw-dev changed the base branch from release-2.8 to master July 1, 2023 16:02

apply on top of master

e48bf27

Signed-off-by: Alexandre Gaudreault <alexandre.gaudreault@logmein.com>

agaudreault force-pushed the fix-deadlock branch from 03300c6 to e48bf27 Compare July 1, 2023 18:30

agaudreault mentioned this pull request Jul 1, 2023

feat: add ignoreResourceUpdates to reduce controller CPU usage (#13534) #13912

Merged

13 tasks

jannfis approved these changes Jul 1, 2023

View reviewed changes

jannfis merged commit 53db27e into argoproj:master Jul 1, 2023

gcp-cherry-pick-bot bot mentioned this pull request Jul 1, 2023

fix: deadlock in controller (cherry-pick #14304) #14306

Merged

gcp-cherry-pick-bot bot pushed a commit that referenced this pull request Jul 1, 2023

fix: deadlock in controller (#14304)

8b520a5

Signed-off-by: Alexandre Gaudreault <alexandre.gaudreault@logmein.com>

crenshaw-dev pushed a commit that referenced this pull request Jul 1, 2023

fix: deadlock in controller (#14304) (#14306)

d11142e

Signed-off-by: Alexandre Gaudreault <alexandre.gaudreault@logmein.com> Co-authored-by: Alexandre Gaudreault <alexandre.gaudreault@logmein.com>

yyzxw pushed a commit to yyzxw/argo-cd that referenced this pull request Aug 9, 2023

fix: deadlock in controller (argoproj#14304)

83913cb

Signed-off-by: Alexandre Gaudreault <alexandre.gaudreault@logmein.com>

suzaku mentioned this pull request Aug 15, 2023

use std lib suzaku/argo-cd#2

Closed

tesla59 pushed a commit to tesla59/argo-cd that referenced this pull request Dec 16, 2023

fix: deadlock in controller (argoproj#14304)

208352a

Signed-off-by: Alexandre Gaudreault <alexandre.gaudreault@logmein.com>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix: deadlock in controller #14304

fix: deadlock in controller #14304

agaudreault commented Jul 1, 2023 •

edited

Loading

codecov bot commented Jul 1, 2023 •

edited

Loading

jannfis commented Jul 1, 2023

crenshaw-dev commented Jul 1, 2023

crenshaw-dev commented Jul 1, 2023

jannfis left a comment

fix: deadlock in controller #14304

fix: deadlock in controller #14304

Conversation

agaudreault commented Jul 1, 2023 • edited Loading

codecov bot commented Jul 1, 2023 • edited Loading

Codecov Report

jannfis commented Jul 1, 2023

crenshaw-dev commented Jul 1, 2023

crenshaw-dev commented Jul 1, 2023

jannfis left a comment

Choose a reason for hiding this comment

agaudreault commented Jul 1, 2023 •

edited

Loading

codecov bot commented Jul 1, 2023 •

edited

Loading