[SSL: UNEXPECTED_EOF_WHILE_READING] EOF occurred in violation of protocol (_ssl.c:1007) #77
Comments
Mark-Simulacrum
added a commit
to Mark-Simulacrum/promote-release
that referenced
this issue
Oct 23, 2023
This is likely to be very verbose, but hopefully will eventually give us request IDs for rust-lang#77 which we can ask about to S3, or give us enough information to fix it locally. This has broken 2 nightlies in 12 days, which is a pretty high incidence rate so far.
Mark-Simulacrum
added a commit
that referenced
this issue
Nov 5, 2023
This is likely to be very verbose, but hopefully will eventually give us request IDs for #77 which we can ask about to S3, or give us enough information to fix it locally. This has broken 2 nightlies in 12 days, which is a pretty high incidence rate so far.
|
We disabled debug logging because it was rather excessive, but this happened again today: We might want to consider an in-code retry of all aws s3 commands or something... |
|
Happened again today: |
|
I manually enabled S3 access logs for static-rust-lang-org. My hope is that gives us request IDs without changing the client, but given the error is an SSL error I'm not super optimistic it'll be tied to any particular request. |
|
Failed again today: Looking at S3 logs (now that we have those), I'm not seeing any references to this file, though there are uploads around the same time from the codebuild job. I don't see any references to this particular file path or obvious errors for PUT requests. So unfortunately no real luck from having those logs... |
|
Occurred again today: |
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
We have seen this error several times recently:
on:
See below for a capture from the most recent occurence (2023-10-23). This appears to be a spurious error, unfortunately we don't yet have request IDs or similar to investigate further, and depending on when the failure occurs they may not be available. This may be due to new client software (e.g., awscli) or due to upstream changes, it's not clear.
Both failures happened using
awscli 1.22.34-1from Ubuntu 22.04 archives, starting just after we landed 19ed30b (which rebuilt the docker image). But we were using that awscli previously as well.I'm not sure whether underlying openssl versions or similar changed across that time range.
The text was updated successfully, but these errors were encountered: