harvester: [Doc] upgrade stuck while upgrading system service with alertmanager and prometheus

Describe the bug

Perform 3 nodes upgrade, but it stuck in Upgrading System Service for hours

To Reproduce

Steps to reproduce the behavior:

Install Harvester with any nodes having additional ports for VLAN
Setup Cluster Network/Config with additional port
Setup storage-network with the cluster network
Remove storage-network and Cluster Network
Perform upgrade by click the Upgrade button in Dashboard

Expected behavior

Upgrade should sccess.

Environment:

Harvester ISO version: v1.1.1 to v1.1.2
ui-source Option: Auto
Underlying Infrastructure (e.g. Baremetal with Dell PowerEdge R630): Baremetal DL360 3 nodes

Additional context

It looks similar to #3675 but that workaround won’t work.

Support bundle

supportbundle_685deba0-d009-4dde-aff9-7c9210eeac82_2023-04-27T13-46-26Z.zip

About this issue

Original URL
State: closed
Created a year ago
Comments: 28 (17 by maintainers)

Most upvoted comments

If you have hit this issue, please use following workaround to patch on the fly (it will take around 2 minutes to eliminate the warning message):

kubectl edit managedchart -n fleet-local rancher-monitoring

add following

    - apiVersion: monitoring.coreos.com/v1
      jsonPointers:
      - /spec
      - /status
      kind: Prometheus
      name: rancher-monitoring-prometheus
    - apiVersion: monitoring.coreos.com/v1
      jsonPointers:
      - /spec
      - /status
      kind: Alertmanager
      name: rancher-monitoring-alertmanager

under:

spec:
  diff:
    comparePatches:

w13915984028 on May 26, 2023

@bk201 Check https://github.com/harvester/harvester/files/11388765/supportbundle_fe9a860e-fe6b-441b-a237-6ed80bf6f605_2023-05-03T18-02-11Z.zip - it’s from a different issue, but should display the monitoring resources being in an undesired state after 1.1.2

oivindoh on May 4, 2023