Details
-
Bug
-
Resolution: Fixed
-
Critical
-
6.6.3, 6.6.4, 6.6.5
-
Untriaged
-
0
-
Unknown
Description
Problem
Diag log is my goto to build a quick timeline of the cluster:
rebalance,
failovers
restarts
Unfortunately if the cluster has metadata using a large amount of memory the log is just full of:
Metadata overhead warning. Over 73% of RAM allocated to bucket
|
Basically the log ends up being a few hour and totally useless.
This really impact the ability to debug problems at speed.