I've tried increasing various system limits (using ulimit), generally doubling
and increasing 10-fold any value. None of the following has made any difference:
# -n : The maximum number of open file descriptors (most systems do not allow
this value to be set)ulimit -n 2048# -u : The maximum number of processes
available to a single userulimit -u 5149520# -s : The maximum stack sizeulimit
-s unlimited# -l : Maximum locked memoryulimit -l 128# -i : The maximum number
of pending signalsulimit -i 5149520# -q : POSIX message queuesulimit -q 8192000
Looking in /proc/<pid>/limits I've confirmed the limits have been raised
from:Limit Soft Limit Hard Limit
UnitsMax cpu time unlimited unlimited
secondsMax file size unlimited unlimited
bytesMax data size unlimited unlimited
bytesMax stack size 10485760 unlimited
bytesMax core file size 0 0
bytesMax resident set unlimited unlimited
bytesMax processes 1024 514952
processesMax open files 1024 1024
filesMax locked memory 65536 65536
bytesMax address space unlimited unlimited
bytesMax file locks unlimited unlimited
locksMax pending signals 514952 514952
signalsMax msgqueue size 819200 819200
bytesMax nice priority 0 0Max realtime priority
0 0Max realtime timeout unlimited unlimited
us
to:Limit Soft Limit Hard Limit UnitsMax
cpu time unlimited unlimited secondsMax file
size unlimited unlimited bytesMax data size
unlimited unlimited bytesMax stack size
unlimited unlimited bytesMax core file size 0
0 bytesMax resident set unlimited
unlimited bytesMax processes 5149520
5149520 processesMax open files 2048
2048 filesMax locked memory 131072 131072
bytesMax address space unlimited unlimited
bytesMax file locks unlimited unlimited
locksMax pending signals 5149520 5149520
signalsMax msgqueue size 8192000 8192000
bytesMax nice priority 0 0Max realtime priority
0 0Max realtime timeout unlimited unlimited
us
I have 166 virtual hosts, and 124 proxy balancers. As soon as I try to increase
the number of proxy balancers to 125 I get the segmentation fault error. I only
have one set of log files (1 access log, 1 error log), I don't have separate
log files for each vhost.
Thanks,
Paul
From: paul_beck...@outlook.com
To: users@httpd.apache.org
Date: Fri, 25 Jul 2014 19:00:44 +0100
Subject: [users@httpd] Segmentation Fault - too many proxy balancers
My apache server has started segmentation faulting all the time (seems to log a
segmentation fault every few requests to the apache error log):
[Fri Jul 25 06:25:42.046752
2014] [core:notice] [pid 11226:tid 140006078953216] AH00052: child pid 11715
exit signal Segmentation fault (11)
This appears to be due to the number of proxy balancers I have configured
(problem isn't related to any one specific proxy balancer, adding / removing
any of the proxy balancers causes the problem to appear/disappear). I'm using
Apache HTTPD as a reverse proxy for a lot of load-balanced (by apache httpd)
application servers. My googling so far hasn't found any specific limit on the
number of proxy, or how I can increase this.
I am running Apache HTTPD 2.4.9, built from source on RHEL6.
I would be very grateful if anyone can shed more light on this, and assuming
I'm right about a limit: point my in the right direction as to how I can
increase this.
Thanks,
Paul
