Origin PR: https://github.com/openshift/origin/pull/18514 Upstream PR: https://github.com/kubernetes/kubernetes/pull/59506 The reason why this error was occurring was due to the large amount of events that were being generated. The general issue, however, applies to any resource whose quantity exceeds a given chunk-size (defaults to "500" for `oc get`). When the total number of a requested resource exceeds the request's chunk-size, multiple requests end up being made to the same endpoint, with each request producing its own "Info" object. Although we are dealing with a single resource kind, `oc get --watch` produces an error as it associates more than one "Info" with more than one resource kind. The proposed patch modifies the `oc get` command to only fail if more than one "Info" is returned AND at least one of the infos has different group-version-kind information. This bug can be reproduced locally by creating some resources on a cluster, and running "oc get -w" with a small chunk size value: ``` $ oc get events -w --chunk-size=2 error: watch is only supported on individual resources and resource collections - 17 resources were found ```
Upstream PR [1] has merged. Waiting on Origin [2]. 1. https://github.com/kubernetes/kubernetes/pull/59506 2. https://github.com/openshift/origin/pull/18514
$ oc get events -w [--chunk-size=2] does not reproduced on oc v3.9.0-0.47.0
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-2018:1816