-
-
Notifications
You must be signed in to change notification settings - Fork 307
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
'Decoder reached invalid state' in underlying Reqwest client #144
Comments
According to Nemo157's comment Nullus157/async-compression#79 (comment) he believes the issue is in this crate, particularly in this loop: https://github.com/clux/kube-rs/blob/6fc2fc70903f46c69d8239ade8b3266b3a66ef95/src/client/mod.rs#L165 |
There may be an improvement in the Stream implementation for |
Oh, thank you. That is interesting. Regardless of async-compression's thing, it's possible that a less brittle way of unfolding herein is using |
Yeah, the only downside is waiting for |
From reading his comment, it sounded like we could pin async-compression as long as we had 0.10.2. But if not, then yeah, that could be a long time to wait. |
That's true! reqwest 0.10.2 uses |
Did some preliminary work in #145 - it was rather involved, and still haven't understood all of it, so going to leave it for a bit. Have a busy week unfortunately. |
Actually did some more stuff just to see how it works in a branch. Feel free to try it out. I am seeing a lot more Reflector restarts when getting EOFs from Reqwest, might be other errors now that we're propagating on. |
I'll let you continue digging into this if you feel it is worth while. Thanks for your work! |
Yeah, will make a release late tonight. Have fixed another issue in #146 |
Released the bump in 0.26.0 via #145. |
I posted this issue in the
async-compression
project as well. I'm seeing a random panic in the async GzipDecoder stream in reqwest. I am uncertain if this is because of the way kube-rs is usingreqwest
or if it's because ofreqwest
, or if it's because of theasync-compression
crate. Having read the code where the panic occurs, it looks like the only way this could occur is ifpoll_next
is called simultaneously from separate threads, i.e. it is not thread-safe. Anyone else seeing this with v0.25.0?The text was updated successfully, but these errors were encountered: