https://bz.apache.org/bugzilla/show_bug.cgi?id=64730
Bug ID: 64730
Summary: Client hangup causes incorrect http status 500 logged
Product: Apache httpd-2
Version: 2.5-HEAD
Hardware: PC
Status: NEW
Severity: normal
Priority: P2
Component: mod_fcgid
Assignee: [email protected]
Reporter: [email protected]
Target Milestone: ---
When a client resets the connection to the server while fcgid is still serving
a request, the following is logged:
[fcgid:warn] [pid 8218] (104)Connection reset by peer: [client a.b.c.d:57986]
mod_fcgid: ap_pass_brigade failed in handle_request_ipc function
In the access log this request is logged as a 500 internal server error. This
is incorrect: the response code is already sent to the client and probably not
a 500.
Steps to reproduce:
Create an fcgi script which does the following:
1) generate a "302 moved" response code, redirecting the client to another url
2) flush the response code to the fcgi daemon. Can be achieved by sending a
very large response body.
3) linger around for a while, don't finish up this request too quickly.
The client will parse the redirect, close the connection and go the the
indicated url.
The closed connection causes the warning in the server error log and the 500
response code in the access log.
What do I expect:
a) don't log a warning. It's not indicative of a situation which requires
attention. Logging as "info" is good enough.
b) log the correct response code in the access log
--
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
