Panic during master startup

[smarterclayton]

On a local openshift start master branch:

E0926 18:15:27.158224   41901 runtime.go:66] Observed a panic: "invalid memory address or nil pointer dereference" (runtime error: invalid memory address or nil pointer dereference)
/Volumes/development/projects/origin/src/github.com/openshift/origin/_output/local/go/src/github.com/openshift/origin/vendor/k8s.io/apimachinery/pkg/util/runtime/runtime.go:72
/Volumes/development/projects/origin/src/github.com/openshift/origin/_output/local/go/src/github.com/openshift/origin/vendor/k8s.io/apimachinery/pkg/util/runtime/runtime.go:65
/Volumes/development/projects/origin/src/github.com/openshift/origin/_output/local/go/src/github.com/openshift/origin/vendor/k8s.io/apimachinery/pkg/util/runtime/runtime.go:51
/usr/local/Cellar/go/1.9/libexec/src/runtime/asm_amd64.s:509
/usr/local/Cellar/go/1.9/libexec/src/runtime/panic.go:491
/usr/local/Cellar/go/1.9/libexec/src/runtime/panic.go:63
/usr/local/Cellar/go/1.9/libexec/src/runtime/signal_unix.go:367
/Volumes/development/projects/origin/src/github.com/openshift/origin/_output/local/go/src/github.com/openshift/origin/vendor/k8s.io/kubernetes/pkg/controller/node/taint_controller.go:288
/Volumes/development/projects/origin/src/github.com/openshift/origin/_output/local/go/src/github.com/openshift/origin/vendor/k8s.io/kubernetes/pkg/controller/node/nodecontroller.go:385
/Volumes/development/projects/origin/src/github.com/openshift/origin/_output/local/go/src/github.com/openshift/origin/vendor/k8s.io/kubernetes/pkg/controller/node/controller_utils.go:329
/Volumes/development/projects/origin/src/github.com/openshift/origin/_output/local/go/src/github.com/openshift/origin/vendor/k8s.io/client-go/tools/cache/controller.go:192
<autogenerated>:1
/Volumes/development/projects/origin/src/github.com/openshift/origin/_output/local/go/src/github.com/openshift/origin/vendor/k8s.io/client-go/tools/cache/shared_informer.go:547
/usr/local/Cellar/go/1.9/libexec/src/runtime/asm_amd64.s:2337
I0926 18:15:27.258365   41901 controller_utils.go:1032] Caches are synced for node controller
I0926 18:15:27.258417   41901 taint_controller.go:182] Starting NoExecuteTaintManager
I0926 18:15:27.258469   41901 nodecontroller.go:542] Initializing eviction metric for zone:

Line of code is

tc.nodeUpdateQueue.Add(updateItemInterface(updateItem))

Is this a data race?

[smarterclayton]

@openshift/sig-master

[ironcladlou]

Looks like kubernetes/kubernetes@928ac58 (see also: kubernetes/kubernetes#51492)

[deads2k]

Looks like kubernetes/kubernetes@928ac58 (see also: kubernetes/kubernetes#51492)

Actually fixed here: kubernetes/kubernetes#51633 ?

[ironcladlou]

@deads2k

Looks like kubernetes/kubernetes@928ac58 (see also: kubernetes/kubernetes#51492)

Probably want both. kubernetes/kubernetes@928ac58 would fix the specific observed panic if the GC started the informers too soon, and kubernetes/kubernetes#51633 would close any remaining gaps.