httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Tobias Lind" <tobl...@hotmail.com>
Subject Fix in 1.3.24 for processes hanging in keepalive and read-state?
Date Fri, 01 Mar 2002 15:09:58 GMT
Hello!
We have been looking forward to v1.3.24 since we have huge problems with 
hanging threads with apache 1.3.23 under heavy load and large POSTs from the 
clients. Is this beeing fixed? I haven't seen anything about it!

This bug has been reported by others also:
http://bugs.apache.org/index.cgi/full/8725
http://bugs.apache.org/index.cgi/full/8261

Here is a snip from our server-status-log yesterday (only showing the 
hanging threads with huge SS):

56-0  26070 0/10481/10481  W 6.89  10612 0    0.0   52.83 52.83 
217.208.48.65   www.kamrat.com POST /kamrat/kamrat?do=doEdit HTTP/1.1
74-0  26100 0/12343/12343  W 8.39  6400  0    0.0   60.04 60.04 213.64.94.13 
    www.kamrat.com POST /kamrat/kamrat?do=doEdit HTTP/1.1
83-0  26134 0/11033/11033  W 7.04  10622 0    0.0   51.66 51.66 
217.208.48.65   www.kamrat.com POST /kamrat/kamrat?do=doEdit HTTP/1.1
86-0  26137 0/10767/10767  W 6.53  10635 0    0.0   51.94 51.94 
217.208.48.65   www.kamrat.com POST /kamrat/kamrat?do=doEdit HTTP/1.1
143-0 2322  0/2044/6171    W 1.28  10624 0    0.0   9.29  30.17 
217.208.48.65   www.kamrat.com POST /kamrat/kamrat?do=doEdit HTTP/1.1
156-0 20893 0/195/4991     W 0.14  10620 0    0.0   1.16  24.70 
217.208.48.65   www.kamrat.com POST /kamrat/kamrat?do=doEdit HTTP/1.1
170-0 811   0/89/3470      W 0.06  10629 0    0.0   0.51  17.10 
217.208.48.65   www.kamrat.com POST /kamrat/kamrat?do=doEdit HTTP/1.1

The operation that were called when the threads got stuck (doEdit) is an 
operation which should receive quite much data (around 5-10 kilobytes) 
posted from the client. Most of the time it's this operation that causes the 
hang (it's the one that receives most data of all our operations)

We have been running our application on a dual PII 400 MHz, Apache 1.3.19, 
Linux 2.2.17-14smp kernel for two years before (under very heavy load), and 
we have never seen hanging threads like this before!

We're currently running RedHat Linux 7.2
Our system is a dual PIII 1266MHz, 2.3 Gb RAM
Linux Ixtapa 2.4.9-21smp #1 SMP Thu Jan 17 14:01:48 EST 2002 i686 unknown
Server Version: Apache/1.3.23 (Unix) ApacheJServ/1.1.2

We _have_ tried the patch suggested in 
http://bugs.apache.org/index.cgi/full/8261
But it doesn't seem to fix/workaround this problem in our system :(

We have also tried the recv_timeout-patch 
(http://www.webcon.net/opensource/apache/)

The Timeout-variable in httpd.conf is set to 180 secs in our system, and the 
RecvTimeout (introduced by the recv_timeout-patch) to 10 seconds

Please take a look at this problem! It's a bad one!

Regards
/Tobias Lind

_________________________________________________________________
Join the world’s largest e-mail service with MSN Hotmail. 
http://www.hotmail.com


Mime
View raw message