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
Describe the bug
Getting a lot of hyper: connection closed before message completed errors which seem to be hyperium/hyper#2136. rust-s3 does not seem to configure any sort of keepalive value, I assume doing that fixes the issue. AWS NodeJS SDK docs indicate a default value of 1000ms for keepalive.
Another way would be to set pool_idle_timeout to match the S3 server value of 20s, according to comments in the hyper issue.
To Reproduce
Open a connection, sleep for more than 20s then try to use said connection.
Expected behavior
Connections remain open and usable.
Environment
Rust version: 1.80.1
lib version: 0.35.1
The text was updated successfully, but these errors were encountered:
Another way would be to set pool_idle_timeout to match the S3 server value of 20s, according to comments in the hyper issue.
This does not seem to hold true. I've tried different pool_idle_timeout durations, as low as 5 seconds, but the error persists, although less frequent.
Even tcp_keepalive does not seem to help.
Oh well, retrying seems to be the solution for now. Seems wasteful though.
Describe the bug
Getting a lot of
hyper: connection closed before message completed
errors which seem to be hyperium/hyper#2136.rust-s3
does not seem to configure any sort ofkeepalive
value, I assume doing that fixes the issue. AWS NodeJS SDK docs indicate a default value of 1000ms forkeepalive
.Another way would be to set
pool_idle_timeout
to match the S3 server value of 20s, according to comments in the hyper issue.To Reproduce
Open a connection, sleep for more than 20s then try to use said connection.
Expected behavior
Connections remain open and usable.
Environment
The text was updated successfully, but these errors were encountered: