Bug 1925614
| Summary: | Error: InstallPlan.operators.coreos.com not found | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | OpenShift Container Platform | Reporter: | Asheth <asheth> | ||||||
| Component: | OLM | Assignee: | Nick Hale <nhale> | ||||||
| OLM sub component: | OperatorHub | QA Contact: | Bruno Andrade <bandrade> | ||||||
| Status: | CLOSED ERRATA | Docs Contact: | |||||||
| Severity: | medium | ||||||||
| Priority: | medium | CC: | agawand, jiazha, krizza, nhale | ||||||
| Version: | 4.6.z | ||||||||
| Target Milestone: | --- | ||||||||
| Target Release: | 4.8.0 | ||||||||
| Hardware: | Unspecified | ||||||||
| OS: | Unspecified | ||||||||
| Whiteboard: | |||||||||
| Fixed In Version: | Doc Type: | Bug Fix | |||||||
| Doc Text: |
Cause: Info log level used for transient errors.
Consequence: Noisy olm-operator logs for the default configuration.
Fix: Change transient error log level to debug.
Result: Less logs noise for the debug configuration.
|
Story Points: | --- | ||||||
| Clone Of: | Environment: | ||||||||
| Last Closed: | 2021-07-27 22:41:35 UTC | Type: | Bug | ||||||
| Regression: | --- | Mount Type: | --- | ||||||
| Documentation: | --- | CRM: | |||||||
| Verified Versions: | Category: | --- | |||||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||||
| Embargoed: | |||||||||
| Attachments: |
|
||||||||
This problem is still there in OCP 4.6.16. From the logs (https://bugzilla.redhat.com/attachment.cgi?id=1755267), it looks like there are two distinct issues of various severity: 1. (low severity) error log when the InstallPlan a Subscription references no longer exists 2. (higher severity) high-frequency reconciliation of CSVs and Operator adoption (where the error log in 1 is generated) We can fairly easily suppress the log in issue 1; issue 2 is very similar to another bug, which has a verified fix that unfortunately hasn't been released to 4.6.z yet according to bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1924257 (In RELEASE_PENDING as of 02/15) I tested today against 4.7 and was unable to trigger the high-frequency reconciliation (i.e. hot-loop). Tomorrow I'll attempt to reproduce against 4.6.16 and then again on the release-4.6 branch of OLM (which already contains the patch). If I can reproduce the hot-loop in 4.6.16 but not release-4.6, it means the two are not related and I'll open a PR with a simple patch to suppress issue 1; if I can (or can't) reproduce the hot-loop in _both_, it means that the root cause is not the same, and I'll need to do more digging. Yesterday I was able to get some similar behavior by following the repro instructions in https://bugzilla.redhat.com/attachment.cgi?id=1755267 and deleting an InstallPlan referenced by a (any) Subscription. Given that the hot-looping patch will (seemingly) be introduced in the next release, I'm reducing the priority/severity of this issue to low and I've opened a PR that will reduce log noise in olm-operator for non-debug settings. Created attachment 1757619 [details]
$ oc get subscription -A -o yaml
Created an operator, delete the install plan, and logs now do not show the messages reported on the Bug. LGTM, marking as VERIFIED. oc exec olm-operator-564748df57-nxzwb -n openshift-operator-lifecycle-manager -- olm --version OCP: 4.8.0-0.nightly-2021-02-18-175130 OLM version: 0.17.0 git commit: ad769c0747d0a7dd5bf501e7425b255535949dc8 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 (Moderate: OpenShift Container Platform 4.8.2 bug fix and security update), 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/RHSA-2021:2438 |
Created attachment 1755263 [details] Output of the command $ oc get subscription -A Description of problem: OLM pod logs repetitively show ""msg":"component not found","error":"InstallPlan.operators.coreos.com \"install-5bwb6\" not found" Version-Release number of selected component (if applicable): 4.6.15 Actual results: OLM pod logs repetitively show ""msg":"component not found","error":"InstallPlan.operators.coreos.com \"install-5bwb6\" not found" Expected results: These error messages should be fixed.