internal/sockstate: Fix sockets ending up in blocking mode when we monitor them for disconnects. #1311
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…nitor them for disconnects. As documented in `os`, calling *os.File.Fd() will put the socket into blocking mode, makes SetDeadline methods stop working, etc. This syscall to SetNonblocking restores desired functionality. This was discovered when testing clustering control plane operations in Dolt, which rely on the timely ability to terminate all client connections by calling Close() on them. We were able to reproduce the issue on macOS by doing the same File() behavior there, despite not having an implementation to actually use the fd for external monitoring of the connection state. To keep as much behavioral parity going forward, I left the fd translation in, since we've observed that it can radically change behavior.
zachmu
approved these changes
Oct 6, 2022
| } | ||
|
|
||
| defer f.Close() | ||
| if isWSL || isProcBlocked { |
There was a problem hiding this comment.
Seems like this check still belongs above the one above, no? Otherwise we get a less useful error message on WSL
I can test this a bit later
There was a problem hiding this comment.
Wait, surely TCPConn.File() on WSL doesn't fail? I was trying to mimic the behavior, same as on macOS, where we ensure the ofile got put back into non-blocking.
There was a problem hiding this comment.
Glad I tested this, it's definitely broken on WSL!
Digging into it, will have some patches a bit later.
There was a problem hiding this comment.
Nevermind! Turns out this was hitting a different unrelated bug, this appears to work fine!
zachmu
approved these changes
Oct 10, 2022
Hydrocharged
deleted the
aaron/fix-monitored-tcpconns-blocking-on-linux
branch
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
As documented in
os, calling *os.File.Fd() will put the socket into blocking mode, makes SetDeadline methods stop working, etc. This syscall to SetNonblocking restores desired functionality.This was discovered when testing clustering control plane operations in Dolt, which rely on the timely ability to terminate all client connections by calling Close() on them.
We were able to reproduce the issue on macOS by doing the same File() behavior there, despite not having an implementation to actually use the fd for external monitoring of the connection state. To keep as much behavioral parity going forward, I left the fd translation in, since we've observed that it can radically change behavior.