Replies: 1 comment
-
Sadly the logging in the placement strategies leaves a lot to be desired :( That said, I would guess that your active volumes per worker is what is limiting you in this case. I would probably make the number of volumes 2-3 times higher than number of containers, as a starting point at least. You usually have a lot more volumes than you do containers. You can see this in the table output from On the Concourse CI system I can see the workers generally have around 7 times the number of volumes compared to containers:
|
Beta Was this translation helpful? Give feedback.
-
Question: How do we find the root cause constraint for the message (and corresponding wait) above?
We have a pretty large concourse deployment with ~40 workers.
Our workers did not have their disks full and were nowhere near the number of containers (max was < 80 and most were
check
,get
, andput
containers)I checked the stdout/stderr logs on the web nodes and could not find anything pointing to an error.
Beta Was this translation helpful? Give feedback.
All reactions