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
Thread killed by timeout reaper when sending large HTTP/2 response #593
Comments
I need a small test case.
And test it with nghttp:
I think this is not your problem. No error is displayed in the we app side. Could you modify this application to reproduce your problem easily? |
I think I can reproduce a similar problem with my nano test case and my experimental client library. Client: my executable at https://github.com/lucasdicioccio/http2-client Server:
If I don't set the timeout with |
I think I found a cause for at least my version of the bug. In @kazu-yamamoto what do you think? |
@lucasdicioccio I think that your observation is right. |
Ack. I'll give it a shot this week-end. |
This patch addresses a bug preventing long HTTP2 connections. See yesodweb#593 for an initial report, as well as yesodweb#593 (comment) for a reproducible example -- which this patch addresses (the actual root cause for yesodweb#593 might differ but seem closely related).
Ping @ip1981 |
It will take some time :) |
Negative. I'm still getting timeout. To be 100% sure I used cabal sandbox, installed warp from the I used 1G file created with Sproxy log:
Did I miss something? |
I finally reproduce @ip1981's problem (with @lucasdicioccio's patch). |
Here is what I did:
What I found is probably a bug of Firefox, not @ip1981's one. |
This patch addresses a bug preventing long HTTP2 connections. See yesodweb#593 for an initial report, as well as yesodweb#593 (comment) for a reproducible example -- which this patch addresses (the actual root cause for yesodweb#593 might differ but seem closely related).
This happens Firefox on macOS only. |
I'm 100% sure that Warp version 3.2.20 fixes this issue. |
With Warp 3.2.8 I'm getting error
Thread killed by Warp's timeout reaper
when sending large (GiB) bodies from backend servers. This looks similar to #351, but happens only on HTTP/2 (HTTP/1.1 to the backend).The docs say
The data is definitely being sent to the client. Also this commit suggests that timeout should be disabled when request body is consumed, that I believe is the case.
Source code:
The text was updated successfully, but these errors were encountered: