Skip to content

Conversation

praveen-influx
Copy link
Contributor

When troubleshooting OOM on WAL replay I noticed unbounded concurrency is leading to the issue. This experimental PR is a way to address it (i.e limit concurrency), by exposing this as a CLI parameter it'll give users a handle to set the limit so that if they run into errors with default, concurrency can be reduced and retried.

This does not guarantee it won't OOM after that when snapshotting as part of WAL replay, but this will take it further than where it fails at the moment.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant