"AIOpsManagement.sh --update" failed with [ERROR] Failed to collect transparent huge pages value

Products

• DX Operational Observability

Issue/Introduction

DX O2 Upgrade AIOpsManagement.sh --update failed with:

[ERROR] Failed to collect transparent huge pages value from <node(s)> 
[ERROR] failed to validate huge pages. See install.log for more details. Fix issues and restart.

Environment

DX O2 24*, 25*

Resolution

Suggestions:

1) As per documentation enable Transparent Huge Pages (THP) Only in Madvise Regions

Run the below command in all nodes:
cat /sys/kernel/mm/transparent_hugepage/enabled

Ensure output is as below:
always (madvise) never

 

2) If above check is in place check for possible network issues

a) Review the install.log and check for possible messages indicating a delay or servers failing to respond:

b) Run some oc/kubctl commands to validate that cluster is responding as expected, for example, in openshift : oc debug node/<node-name>

If the above command takes a long time to respond, it would indicate a network issue

c) It could be just a temporal network issue, you could retry the validation again, in the prompt enter Y again, if you get next prompt, then it would you can safely continue with the upgrade.

Additional Information

Techdocs DX O2 Upgrade