modwsgi

I didn't look at the error log for days, but there has always been a
few IOError tracebacks in my mail daily. I didn't bother to trace the
cause, because most of the time, I assumed it's a client upload timing
out. Now I searched for it, there has been a few bucket brigade lines
in the error log. (note: pid isn't constant)

Yesterday's log:

46x (70007)The timeout specified has expired: mod_wsgi (pid=23537):
Unable to get bucket brigade for request.
43x (104)Connection reset by peer: mod_wsgi (pid=16787): Unable to get
bucket brigade for request.

(Number may seem large, but we handled 197054 POST requests yesterday,
3462660 requests total, it's really not a bad score :P)


Every "problematic" client generates a few of those messages. For us,
it occurs mostly during uploads. For those clients, it takes a few
tries to get a chunk uploaded:

77.246.107.xxx - - [18/Jun/2009:21:51:52 +0200] "POST /client/reports/
live/update/ HTTP/1.1" 500 653 "-" "Jakarta Commons-HttpClient/3.1"
77.246.107.xxx - - [18/Jun/2009:21:52:11 +0200] "POST /client/reports/
live/update/ HTTP/1.1" 500 653 "-" "Jakarta Commons-HttpClient/3.1"
77.246.107.xxx - - [18/Jun/2009:21:52:31 +0200] "POST /client/reports/
live/update/ HTTP/1.1" 500 653 "-" "Jakarta Commons-HttpClient/3.1"
77.246.107.xxx - - [18/Jun/2009:21:52:50 +0200] "POST /client/reports/
live/update/ HTTP/1.1" 200 29 "-" "Jakarta Commons-HttpClient/3.1"


So far, it doesn't look like it does any harm, am I right in assuming
that this is just a issue of not handling the request timeout
correctly?


Miles
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"modwsgi" group.
To post to this group, send email to [EMAIL PROTECTED]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at 
http://groups.google.com/group/modwsgi?hl=en
-~----------~----~----~----~------~----~------~--~---