Skip to content

fix(s3stream): fix the network out over-consumed #2752

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Aug 13, 2025
Merged

Conversation

superhx
Copy link
Collaborator

@superhx superhx commented Aug 12, 2025

Consider the following scenario:

  1. A Fetch request contains partitions P1 and P2. The data of P1 is in LogCache, while the data of P2 is not.
  2. First, a fast read will be attempted. At this time, P1 will return data and consume Network Out, and P2 will return a FastReadException.
  3. Due to the FastReadException, the entire Fetch attempts a slow read. At this time, both P1 and P2 return data and consume Network Out.
  4. At this point, the Network Out in step 2 is consumed repeatedly.

Solution: Move the S3Stream network out consumption to ElasticReplicaManager. Avoid the network out traffic over-consumed, when there are mixin(tail read & catch-up read) partitions reading.

Move the S3Stream network out consumption to ElasticReplicaManager.
Avoid the network out traffic over-consumed, when there are
mixin(tail read & catch-up read) partitions reading.

Signed-off-by: Robin Han <[email protected]>
@superhx superhx merged commit 5520dde into main Aug 13, 2025
7 checks passed
@superhx superhx deleted the fix_over_limit branch August 13, 2025 02:25
superhx added a commit that referenced this pull request Aug 13, 2025
Consider the following scenario:

A Fetch request contains partitions P1 and P2. The data of P1 is in LogCache, while the data of P2 is not.
First, a fast read will be attempted. At this time, P1 will return data and consume Network Out, and P2 will return a FastReadException.
Due to the FastReadException, the entire Fetch attempts a slow read. At this time, both P1 and P2 return data and consume Network Out.
At this point, the Network Out in step 2 is consumed repeatedly.
Solution: Move the S3Stream network out consumption to ElasticReplicaManager. Avoid the network out traffic over-consumed, when there are mixin(tail read & catch-up read) partitions reading.

Signed-off-by: Robin Han <[email protected]>
superhx added a commit that referenced this pull request Aug 13, 2025
Consider the following scenario:
1. A Fetch request contains partitions P1 and P2. The data of P1 is in LogCache, while the data of P2 is not.
2. First, a fast read will be attempted. At this time, P1 will return data and consume Network Out, and P2 will return a FastReadException.
3. Due to the FastReadException, the entire Fetch attempts a slow read. At this time, both P1 and P2 return data and consume Network Out.
4. At this point, the Network Out in step 2 is consumed repeatedly.

Solution: Move the S3Stream network out consumption to ElasticReplicaManager. Avoid the network out traffic over-consumed, when there are mixin(tail read & catch-up read) partitions reading.
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.

2 participants