You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
## Description
Addresses #2593 by
getting rid of the busy loop entirely. The loop was checking to make
sure all values from the channel were read before closing the channel.
When following logs, if a stream is cancelled before reading all logs in
the channel, it loops infinitely. This changes the consumer to loop
until it's read everything from the channel, even after its been closed.
Also, adds a port for pprof to help with profiling in the future.
## Is this change user facing?
YES
## References
#2593
What's your CLI version?
1.4.1
Description & steps to reproduce
Spin up something:
Follow the logs a few times:
At this point everything is fine. But if you stop those processes, each call will use 100% CPU:
At this point the kurtosis-engine process will be using roughly 1000% cpu.
Desired behavior
I guess this loop should either have a delay for some kind of upper bound on the amount of time that it will wait.
https://github.com/kurtosis-tech/kurtosis/blob/v1.4.1/engine/server/engine/centralized_logs/client_implementations/persistent_volume/persistent_volume_logs_database_client.go#L178-L185
Since it's looping forever, I would guess the check on
len(logsChan)
isn't doing what we expect.What is the severity of this bug?
Papercut; this bug is frustrating, but I have a workaround.
What area of the product does this pertain to?
CLI: the Command Line Interface
The text was updated successfully, but these errors were encountered: