-
Bug
-
Resolution: Unresolved
-
Normal
-
None
-
4.17
-
None
-
Moderate
-
None
-
False
-
Description of problem:
When MetalLB is installed in an OCP cluster the must-gather collection can become very large due to the collection of full frr logs. This makes difficult for customer to collect must-gather for some clusters. KB 7068652 has been written to manually disable the collection of those logs to help customers. Here I'm proposing to disable the collection of full frr logs from the must-gather script, so only the latest lines would be collected. In case it would be needed, the support can ask to retrieve the full logs from the pod. A pull request is already in place.
Version-Release number of selected component (if applicable):
How reproducible:
By installing MetalLB in FRR mode, let it run until logs become very large and then try to collect a must-gather
Steps to Reproduce:
1. Install MetalLB in FRR mode 2. Let it run for the needed time for FRR logs to grow 3. Try to collect a must-gather
Actual results:
A too large must-gather collection (hundreds of GB)
Expected results:
A must-gather collection of a reasonable size (no more than 30GB for a 10-node cluster)
Additional info:
There's already a pull request to the must-gather repo to solve this issue.