How to set Kafka log.retention.hours with Operator? #9812
Replies: 2 comments 12 replies
-
I guess you would need to check the logs to see why is not rolling the brokers. You might find useful information in the |
Beta Was this translation helpful? Give feedback.
-
Ultimately looks like somehow the When the kafka pods restarted they became decoupled or similar from the Updating the disk size in the Root cause is still unknown though, we didn't touch the |
Beta Was this translation helpful? Give feedback.
-
Hi, we have a Strimzi jaeger kafka cluster in k8s, version 3.4.0 it looks like. Our lead kafka box is running low on disk on it's pvc where everything gets stored.
In the past, having noted that the
log.retention.hours
was showing as168
on startup, I pruned that down by editing thekafka
resource and adding:To reduce retention from 7 days to 4. When I did this previously the disk was almost entirely full, and editing and saving that change initiated a restart of the pods and they began to prune down on their own.
Now some months later we're hit with the same situation but I see the default
168
hours set in again, and traces go back 7 full days. When I repeat the same steps and edit thekafka
resource, adding the samelog.retention.hours: 96
configuration line, which wasn't present when I checked this time, nothing happens. No restarts, no pruning, no log activity on the pods at all.I'm unsure if any other changes occurred on the setup since the first time some months ago, but I doubt it. When I check the
/tmp/strimzi.properties
file on the pods it doesn't have anything set forlog.retention.hours
so it must be using defaults...but why when I have it set?How does one set this value to less than the default, correctly, if not via the
kafka
resource?Beta Was this translation helpful? Give feedback.
All reactions