RHCOS aarch64 node scaling failing after upgrade to 4.19
Issue
When scaling up new ARM 64 (aarch64) nodes in an OpenShift 4.19 cluster or later, if that cluster was originally installed with OCP 4.12 or earlier, the GRUB bootloader will not be able to load the kernel. This is because it uses a format that is not supported by older GRUB versions. The error message that is observed when on an affected system will look like:
error: ../../grub-core/loader/arm64/linux.c:58:invalid magic number.
error: ../../grub-core/loader/arm64/linux.c:278:you need to load the kernel first.
Press any key to continue...
This issue does not affect any architectures other than ARM 64 (aarch64).
Existing nodes that update to 4.18.11 or later before updating to 4.19 will have their boot loader updated to avoid this issue.
Environment
- OpenShift Container Platform 4.19 ARM 64 cluster nodes that scale up with a RHEL 8 boot image; i.e. the install version of the cluster was 4.12 or earlier. Because there were limited releases where ARM 64 was supported where RHEL8 was the base this should only affect a few platforms:
- Bare Metal - 4.10, 4.11, 4.12
- AWS - 4.11, 4.12
- Azure - 4.12
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.
