Bug 1773870
| Summary: | network operator crashes when infra.status.platformStatus is missing | |||
|---|---|---|---|---|
| Product: | OpenShift Container Platform | Reporter: | Chet Hosey <ChetRHosey> | |
| Component: | Networking | Assignee: | Ricardo Carrillo Cruz <ricarril> | |
| Networking sub component: | openshift-sdn | QA Contact: | zhaozhanqi <zzhao> | |
| Status: | CLOSED ERRATA | Docs Contact: | ||
| Severity: | urgent | |||
| Priority: | urgent | CC: | arghosh, ccoleman, jmalde, nagrawal, openshift-bugs-escalate, ricarril, ssadhale, svaughn, trees | |
| Version: | 4.3.0 | Flags: | ricarril:
needinfo-
ricarril: needinfo- ricarril: needinfo- |
|
| Target Milestone: | --- | |||
| Target Release: | 4.4.0 | |||
| Hardware: | Unspecified | |||
| OS: | Unspecified | |||
| Whiteboard: | ||||
| Fixed In Version: | Doc Type: | If docs needed, set a value | ||
| Doc Text: | Story Points: | --- | ||
| Clone Of: | ||||
| : | 1778235 1781558 1787765 (view as bug list) | Environment: | ||
| Last Closed: | 2020-05-13 21:52:51 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: | ||||
| Bug Depends On: | ||||
| Bug Blocks: | 1778235, 1781558, 1787765 | |||
|
Description
Chet Hosey
2019-11-19 08:05:39 UTC
https://github.com/openshift/cluster-network-operator/blob/release-4.2/vendor/github.com/openshift/api/config/v1/types_infrastructure.go mentions that the InfrastructureStatus.platform field has been deprecated in favor of InfrastructureStatus.platformStatus.type. Given that platformstatus is noted as 'optional', openshift-network-operator should handle the case where it's missing. In the meantime it's not clear as a user if I *should* be modifying status fields on the infrastructure/cluster CR instance, because usually status fields are hands-off from a user standpoint. As you say, status should be treated as read-only. Is there a reason why you are trying to modify them? I'm looking at manuals if there's something in docs for some sort of procedure that has that but not finding it. Can you comment? This is answered by the name of this bug: “network operator crashes when infra.status.platformStatus is missing”. The network operator crashes without a value there. It is unable to apply proxy settings. Yikes, missed that, thanks. I will try to repro and work on a fix. Much appreciated. I’m happy to provide additional info if there’s any that might help. A fresh 4.2 install seems to include the value; a cluster upgraded 4.1.23 -> 4.2.4 did not. The network operator crash relates to applying proxy settings. And that’s the motivation behind wanting to go 4.1 -> 4.2 so aggressively. We don’t have an option to allow direct internet access. Remote IPs for CDNs change often so we have frequent issues pulling container images. So we don’t want to put critical apps on OCP without proxy support, but our existing prod cluster is 4.1. FYI I've also filed https://bugzilla.redhat.com/show_bug.cgi?id=1779299 against the CVO. I'm doubtful that new features are extensively tested on upgraded clusters, as opposed to freshly built clusters. I'd expect either the CVO or a responsible operator to populate the new field on upgrades to ease the potential burden on other components, like the network operator, to include such test cases. The workaround should be fine. FWIW, the fix landed in master, it is pending being picked up by QE. Also backported to 4.2 (will also backport to 4.3, it seems we already branched). 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-2020:0581 |