Bug 1843944 - openshift-kube-proxy pods are crashing when upgrading from 4.3.8 to 4.3.10
Summary: openshift-kube-proxy pods are crashing when upgrading from 4.3.8 to 4.3.10
Keywords:
Status: CLOSED NOTABUG
Alias: None
Product: OpenShift Container Platform
Classification: Red Hat
Component: Networking
Version: 4.3.z
Hardware: Unspecified
OS: Unspecified
high
high
Target Milestone: ---
: 4.3.z
Assignee: Alexander Constantinescu
QA Contact: zhaozhanqi
URL:
Whiteboard:
Depends On: 1842673 1843940 1843943
Blocks:
TreeView+ depends on / blocked
 
Reported: 2020-06-04 13:15 UTC by Ben Bennett
Modified: 2020-09-03 17:21 UTC (History)
4 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of: 1843943
Environment:
Last Closed: 2020-09-03 17:21:20 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)

Description Ben Bennett 2020-06-04 13:15:32 UTC
+++ This bug was initially created as a clone of Bug #1843943 +++

+++ This bug was initially created as a clone of Bug #1843940 +++

+++ This bug was initially created as a clone of Bug #1842673 +++

Description of problem: Cluster is installed with 4.3.8 and because we are using a 3rd party network plugin and when upgrading to 4.3.10 we are blocked because of below  bug :

https://bugzilla.redhat.com/show_bug.cgi?id=1820778 

T

Version-Release number of selected component (if applicable):


How reproducible:

Install 4.3.8 with custom network plugin. try to upgrade to 4.3.10  but kube-proxy pods are in crashloopbackoff state. Error shows as below :

  finishedAt: "2020-05-29T07:35:12Z"
        message: "ut mounting /lib/modules\nW0529 07:33:42.293988       1 proxier.go:597]
          Failed to load kernel module ip_vs with modprobe. You can ignore this message
          when kube-proxy is running inside container without mounting /lib/modules\nW0529
          07:33:42.294780       1 proxier.go:597] Failed to load kernel module ip_vs_rr
          with modprobe. You can ignore this message when kube-proxy is running inside
          container without mounting /lib/modules\nW0529 07:33:42.295505       1 proxier.go:597]
          Failed to load kernel module ip_vs_wrr with modprobe. You can ignore this
          message when kube-proxy is running inside container without mounting /lib/modules\nW0529
          07:33:42.296201       1 proxier.go:597] Failed to load kernel module ip_vs_sh
          with modprobe. You can ignore this message when kube-proxy is running inside
          container without mounting /lib/modules\nW0529 07:33:42.296933       1 proxier.go:597]
          Failed to load kernel module nf_conntrack_ipv4 with modprobe. You can ignore
          this message when kube-proxy is running inside container without mounting
          /lib/modules\nI0529 07:33:42.297182       1 server.go:494] Neither kubeconfig
          file nor master URL was specified. Falling back to in-cluster config.\nI0529
          07:33:42.313013       1 node.go:135] Successfully retrieved node IP: 172.22.104.26\nI0529
          07:33:42.313039       1 server_others.go:146] Using iptables Proxier.\nI0529
          07:33:42.313332       1 server.go:529] Version: v0.0.0-master+$Format:%h$\nI0529
          07:33:42.314087       1 conntrack.go:52] Setting nf_conntrack_max to 131072\nI0529
          07:33:42.314445       1 config.go:313] Starting service config controller\nI0529
          07:33:42.314459       1 shared_informer.go:197] Waiting for caches to sync
          for service config\nI0529 07:33:42.314986       1 config.go:131] Starting
          endpoints config controller\nI0529 07:33:42.315000       1 shared_informer.go:197]
          Waiting for caches to sync for endpoints config\nI0529 07:33:42.414678       1
          shared_informer.go:204] Caches are synced for service config \nI0529 07:33:42.415117
                                                                                                                                179,3         89%
reason: Error
        startedAt: "2020-05-29T07:33:42Z"
    name: kube-proxy
    ready: false
    restartCount: 334
    started: false
    state:
      waiting:
        message: back-off 5m0s restarting failed container=kube-proxy pod=openshift-kube-proxy-2h7tw_openshift-kube-proxy(be11c6f8-1b91-4917-8605-f85be7012b9f)
        reason: CrashLoopBackOff




Steps to Reproduce:
1.
2.
3.

Actual results: Failed to upgrade to 4.3.10


Expected results: Should have succeeded with 3rd party network plugin.


Additional info:

Blocker bug : https://bugzilla.redhat.com/show_bug.cgi?id=1820778

--- Additional comment from Ben Bennett on 2020-06-04 09:07:00 EDT ---

We need to document the procedure to upgrade this:
- Something like disabling CNO so that you can disable the readiness probe on kube-proxy so that you can do the upgrade

Unless there is another way to force the upgrade.

--- Additional comment from Ben Bennett on 2020-06-04 09:08:35 EDT ---

Note that this is not an issue in anything but 4.3.8, so duplicating it, and closing this for releases other than 4.3.

Comment 1 Ben Bennett 2020-09-03 17:21:20 UTC
Closing.  This seems to have been resolved by https://bugzilla.redhat.com/show_bug.cgi?id=1820778


Note You need to log in before you can comment on or make changes to this bug.