Bug 1775849

(In reply to Eric Matysek from comment #0)
> Description of problem:
> NFD pods disappear after cluster upgrade
> upgrading from 4.2.7 to 4.3.0-0.nightly-2019-11-21-122827
> 
> Version-Release number of selected component (if applicable):
> 
> 
> How reproducible:
> 100%
> 
> 
> Steps to Reproduce:
> 1. Deploy 4.2.7 cluster
> 2. git clone https://github.com/openshift/cluster-nfd-operator
> 3. make deploy
> 4. oc adm upgrade --to-image
> registry.svc.ci.openshift.org/ocp/release:4.3.0-0.nightly-2019-11-21-122827
> --force --allow-explicit-upgrade
> 
> Actual results:
> before upgrade:
> [ematysek@jump ~]$ oc get all -n openshift-nfd
> NAME                   READY   STATUS    RESTARTS   AGE
> pod/nfd-master-gvj7n   1/1     Running   0          32s
> pod/nfd-master-kmtx8   1/1     Running   0          32s
> pod/nfd-master-wg47s   1/1     Running   0          32s
> pod/nfd-worker-csdwm   1/1     Running   2          33s
> pod/nfd-worker-nxqrv   1/1     Running   2          33s
> pod/nfd-worker-qnjxn   1/1     Running   2          33s
> 
> after upgrade:
> [ematysek@jump cluster-nfd-operator]$ oc get all -n openshift-nfd
> No resources found in openshift-nfd namespace.
> 
> 
> Expected results:
> NFD pods should still exist
> 
> Additional info:

You were testing the master against a specific openshift release. Master will not always work. Please install NFD from operatorhub in ocp 4.2 and try the upgrade path.

Verified this works with the NFD version in public OperatorHub as well

And by works I mean the bug is present... sorry for the typo

Can you provide any logs or events?

oc get events -n openshift-nfd
oc get events -n openshift-nfd-operator

Is the operator still running?

oc logs -f <operator-name> -n openshift-nfd-operator

What is the status of the DaemonSets in openshift-nfd

oc describe ds -n openshift-nfd

Can you also try to deploy NFD via OLM, and then do the upgrade? 

OLM is responsible for updating the day 2 operators, manually deploying means also manually updating.

This might fix the issue, https://github.com/openshift/cluster-nfd-operator/pull/45

Verification blocked by https://bugzilla.redhat.com/show_bug.cgi?id=1778904

PR 49 is not merged with release-4.2 branch, so it has no effect on a 4.2.x cluster

https://github.com/openshift/cluster-nfd-operator/pull/51 for release-4.2

I don't know why this BZ is talking about 4.2.  This is a 4.3 BZ. I see that this fix is supposedly merged and that 1778904 is VERIFIED. So I'm moving this back ON_QA.

Upgraded successfully without nfd pods disappearing!

[ematysek@jump ~]$ oc get clusterversion
NAME      VERSION                             AVAILABLE   PROGRESSING   SINCE   STATUS
version   4.3.0-0.nightly-2020-01-08-181129   True        False         13m     Cluster version is 4.3.0-0.nightly-2020-01-08-181129

[ematysek@jump ~]$ oc get pods
NAME               READY   STATUS    RESTARTS   AGE
nfd-master-cb57h   1/1     Running   0          14m
nfd-master-pv9c7   1/1     Running   0          14m
nfd-master-rs8tz   1/1     Running   0          14m
nfd-worker-khrhj   1/1     Running   2          14m
nfd-worker-wnwlv   1/1     Running   2          14m
nfd-worker-wxzq7   1/1     Running   2          14m

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHBA-2020:0062