Hey everyone,
After upgrading the RKE version from...
# rker
refined-area-45178
03/18/2025, 3:43 PMHey everyone,
After upgrading the RKE version from 1.3.1 to 1.6.7, and consequently the Kubernetes version from 1.21 to 1.27.16, the deployments have stopped working and are not creating new replicas of pods. Is this a known issue, and does anyone know how to solve it?
refined-area-45178
03/18/2025, 3:45 PMBTW, the system pods works well.
refined-area-45178
03/18/2025, 3:51 PMThe already running PODs of deployments are in running state but after restart of a deployment the new pods are not created even with scale down replicas to 0 and then to an X
c
creamy-pencil-82913
03/18/2025, 4:01 PMdid you go directly from 1.21 to 1.27? You’re supposed to step through 1.22, 1.23, and so on.
r
refined-area-45178
03/18/2025, 5:28 PMYes, I upgraded from 1.21 to 1.27, and everything went smoothly. The cluster is up and running. I identified and fixed the issue, which was related to Consul rather than RKE. After fixing Consul, everything is perfect.
Thanks for your response!
c
creamy-pencil-82913
03/18/2025, 5:33 PMhope that works out. You have probably missed some important resource version migrations. In the future you should refer to https://kubernetes.io/releases/version-skew-policy/#supported-component-upgrade-order
Project policies for API deprecation and API change guidelines requireto not skip minor versions when upgrading, even in single-instance clusters.kube-apiserver
🙌 1
creamy-pencil-82913
03/18/2025, 5:33 PMYou may at some point find that things are more broken than you think because you have skipped MANY minor versions.
🙌 1
r
refined-area-45178
03/19/2025, 9:05 AMI noticed that when a node have 2 IP addresses Pub and private a nodeport does not work on the Public IPs but it is OK for the private ones. Doe you aware of what can be the reason and solution?
C.C @creamy-pencil-82913
refined-area-45178
03/19/2025, 1:22 PMThe above posted issue ^^ has been already resolved.
4 Views