Hide Forgot
Document URL: [1] https://docs.openshift.com/container-platform/3.3/install_config/upgrading/automated_upgrades.html#preparing-for-an-automated-upgrade [2] https://docs.openshift.com/container-platform/3.3/install_config/upgrading/manual_upgrades.html#preparing-for-a-manual-upgrade Section Number and Name: Describe the issue: In manual upgrade to OSCP 3.3 is written to upgrade the kernel, docs [2].. However, the automated installation [1] doesn't mention anything about kernel upgrade. Is it done with upgrade of atomic-openshift-utils? Suggestions for improvement: Add section about kernel upgrade in automated upgrade section in docs. Additional information:
The upgrade of atomic-openshift-utils won't upgrade any kernel. The question is, who is upgrading the kernel of RHEL7 nodes (non atomic ones) when using automated upgrade scripts, the customer or the upgrade scripts. For the manual procedure it is clearly stated in the instructions that the kernel must be updated using yum update kernel. The same is not mentioned in the automated procedure, so we assumed, that this will be done by the upgrade script. So either, please: - update automated installation instructions to clearly state that the customer should first upgrade the kernel by using yum update kernel or - update the upgrade scripts to also update the kernel of RHEL7 nodes during the upgrade
I believe the docs need to be updated to state that the customer should upgrade the kernel themselves manually first. Was trying to double-check myself w/ a test upgrade but am hitting some etcd problems... I couldn't find a spot in openshift-ansible that did this automatically, and checked w/ dgoodwin on the Productization team he did not believe the latest scripts did so either. I'll update again when my upgrade finishes though.
Resolved the etcd upgrade issues. So, after successfully updating my OCP 3.3 cluster to 3.4 using the automated playbook, I can confirm that the kernel was not updated automatically. I still have a pending kernel update available.