CA-401075: remove misleading logs from HTTP client #6158
Merged
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
Recent changes to Http_client were made to log errors and backtraces in
case of unusual errors, in order to better diagnose problems.
Unfortunately, this also led to benign exceptions being logged, which
caused people to suspect problems where the exception was actually
"expected". In particular the following error started appearing in the
logs many times:
Raised Http_client.Parse_error("Expected initial header []")
The case here is that the client has sent an HTTP request, but the
server disconnects before sending a response.
It turns out that this happens commonly when an external connection is
received by xapi and handed over to another process, such as xcp-rrdd.
The connection's file descriptor is passed to the other process, and the
original HTTP request is forwarded over a local socket. The other process
continues to handle the request and sends responses to the forwarded
socket, but never to the local socket, where xapi is waiting for the
response.
This is a quick change that makes the caller of the receive function
aware that no response was sent, without logging scary errors, to fix
the immediate problem. In addition, we should enhance the handover
process to actually send responses, although funcionally this is not an
issue.
Signed-off-by: Rob Hoes <[email protected]>
edwintorok
approved these changes
Dec 4, 2024
| @@ -119,6 +119,8 @@ let response_of_fd_exn_slow fd = | |||
| ; additional_headers= !headers | |||
| ; body= None | |||
| } | |||
| | [] -> | |||
| raise End_of_file | |||
There was a problem hiding this comment.
This could also be raise_notrace if we only use this for control-flow and don't want to log a stacktrace here. (raise_notrace is slightly faster because it doesn't capture a backtrace, whereas raise does (even if we don't query it log it))
psafont
approved these changes
Dec 4, 2024
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.
Recent changes to Http_client were made to log errors and backtraces in case of unusual errors, in order to better diagnose problems. Unfortunately, this also led to benign exceptions being logged, which caused people to suspect problems where the exception was actually "expected". In particular the following error started appearing in the logs many times:
The case here is that the client has sent an HTTP request, but the server disconnects before sending a response.
It turns out that this happens commonly when an external connection is received by xapi and handed over to another process, such as xcp-rrdd. The connection's file descriptor is passed to the other process, and the original HTTP request is forwarded over a local socket. The other process continues to handle the request and sends responses to the forwarded socket, but never to the local socket, where xapi is waiting for the response.
This is a quick change that makes the caller of the receive function aware that no response was sent, without logging scary errors, to fix the immediate problem. In addition, we should enhance the handover process to actually send responses, although funcionally this is not an issue.