upstream deployment: olm pod keeps crashing

[smanpathak]

I followed the install guide to deploy upstream version of olm.

kubectl create -f deploy/upstream/manifests/latest/

Upon inspecting olm namespace, I noticed that one of the pods (registry?) keeps crashing

k -n olm get po  -w
NAME                                READY   STATUS             RESTARTS   AGE
catalog-operator-77ffcc7f88-w8ctc   1/1     Running            0          12s
olm-operator-c77b64c95-57lb9        0/1     CrashLoopBackOff   1          12s
olm-operators-bh6pv                 1/1     Running            0          10s
olm-operator-c77b64c95-57lb9   0/1   Error   2     17s
olm-operator-c77b64c95-57lb9   0/1   CrashLoopBackOff   2     26s
olm-operator-c77b64c95-57lb9   0/1   Error   3     40s
olm-operator-c77b64c95-57lb9   0/1   CrashLoopBackOff   3     41s
olm-operator-c77b64c95-57lb9   0/1   Error   4     89s
olm-operator-c77b64c95-57lb9   0/1   CrashLoopBackOff   4     96s

Looks like a bug

k -n olm logs po/olm-operator-c77b64c95-57lb9
flag provided but not defined: -writeStatusName
Usage of /bin/olm:
  -alsologtostderr
        log to standard error as well as files
  -debug
        use debug log level
  -interval duration
        wake up interval (default 5m0s)
  -kubeconfig string
        absolute path to the kubeconfig file
  -log_backtrace_at value
        when logging hits line file:N, emit a stack trace
  -log_dir string
        If non-empty, write log files in this directory
  -logtostderr
        log to standard error instead of files
  -stderrthreshold value
        logs at or above this threshold go to stderr
  -v value
        log level for V logs
  -version
        displays olm version
  -vmodule value
        comma-separated list of pattern=N settings for file-filtered logging
  -watchedNamespaces -watchedNamespaces=""
        comma separated list of namespaces for olm operator to watch. If not set, or set to the empty string (e.g. -watchedNamespaces=""), olm operator will watch all namespaces in the cluster.

[njhale]

@smanpathak That's actually the olm-operator itself. Looks like there's a mismatch between the deployment manifest and the image it references. We'll get to this soon. Thanks for pointing it out.

[gfenn-newbury]

Can confirm, this is happening for me, too

[ghost]

Same here on fresh minikube running k8s v1.13.5

[jarifibrahim]

On a side note, the quickstart/olm.yaml has the correct image.
User kubectl create -f deploy/upstream/quickstart/olm.yaml to install OLM.

[miguelsorianod]

Hi,

I'm experiencing the same problem. I'm on an OpenShift 3.11 installation and I also see the same error.
In my case I deployed the OKD manifests (and not the upstream ones) located in: deploy/okd/manifests/0.8.1

The contents of the containers of the OLM operator in 0000_50_olm_06-olm-operator.deployment.yaml are:

       containers:
        - name: olm-operator
          command:
          - /bin/olm
          args:
          - -writeStatusName
          - operator-lifecycle-manager
          image: quay.io/coreos/olm@sha256:995a181839f301585a0e115c083619b6d73812c58a8444d7b13b8e407010325f

The SHA256 of quay.io/coreos/olm@sha256:995a181839f301585a0e115c083619b6d73812c58a8444d7b13b8e407010325f corresponds to the 0.8.1 tag in quay.

To discard any problems with Kubernetes/OpenShift I manually downloaded the quay.io/coreos/olm:0.8.1 image and executed the same command and arguments:

msoriano@localhost:~$ docker run -it quay.io/coreos/olm:0.8.1 olm -writeStatusName operator-lifecycle-manager
flag provided but not defined: -writeStatusName
Usage of olm:
  -alsologtostderr
    	log to standard error as well as files
  -debug
    	use debug log level
  -interval duration
    	wake up interval (default 5m0s)
  -kubeconfig string
    	absolute path to the kubeconfig file
  -log_backtrace_at value
    	when logging hits line file:N, emit a stack trace
  -log_dir string
    	If non-empty, write log files in this directory
  -logtostderr
    	log to standard error instead of files
  -stderrthreshold value
    	logs at or above this threshold go to stderr
  -v value
    	log level for V logs
  -version
    	displays olm version
  -vmodule value
    	comma-separated list of pattern=N settings for file-filtered logging
  -watchedNamespaces -watchedNamespaces=""
    	comma separated list of namespaces for olm operator to watch. If not set, or set to the empty string (e.g. -watchedNamespaces=""), olm operator will watch all namespaces in the cluster.

The -writeStatusName flag seems to not be available in that image version.

[ecordell]

We've published a new upstream release which should fix these issues for you.