-
Notifications
You must be signed in to change notification settings - Fork 544
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
Improve http connection reuse #3760
Improve http connection reuse #3760
Conversation
If EOF is not read then the connection is not eligble for reuse. Try to read an extra byte to ensure EOF is read.
The previous value was the MinIO default which is 16 per host and 100 global.
If EOF is not read then the connection is not eligble for reuse. Try to read an extra byte to ensure EOF is read.
be7a0ca
to
7d00c3d
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for working on this.
- Do you think we should put something similar in Azure?
- Can you add a changelog entry detailing the fix
I don't think azure has a similar problem unless it is in the azure library itself.
and this blobClient is from |
what should it be classified in the changelog as? BUGFIX? |
Yeah, I think that's fair. |
Signed-off-by: Joe Elliott <[email protected]>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I went ahead and made the changelog entry b/c of how valuable this fix is.
Thanks!
thanks Joe! |
What this PR does:
Fixes connection reuse for S3 range requests and increases the number of MaxIdleConnsPerHost for the S3 backend to 100 to match the azure settings. Also, assumes GCP has the same connection reuse problem for range requests and performs the same fix. Note: I have tested the fix for S3 range requests and verified it is a problem and the change fixes the problem but I have not verified that it is a problem for GCP or that it fixes the problem for GCP. The GCP HTTP google storage code does look like it effectively returns the body reader so it does look like GCP would have a similar issue.
Which issue(s) this PR fixes:
Fixes #3750
Checklist
CHANGELOG.md
updated - the order of entries should be[CHANGE]
,[FEATURE]
,[ENHANCEMENT]
,[BUGFIX]