httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Chris <chrco...@gmail.com>
Subject [users@httpd] process pile up RUN state
Date Thu, 25 May 2006 06:02:05 GMT
Hi I got a problem on a web server running apache 1.3.35 freebsd 5.4

Every now and then the httpd processes all get stuck on RUN state and
start piling up eventually causing swapping but when things are normal
the server has a lot of free ram and is far from swapping.  2 gig ram
and normal use under 1 gig.

When they are all in RUN state the server is unresponsive on http or
very slow and ssh is fine but a bit slow.  The way out is killall -9
httpd and then it usually runs ok for a while again, this behaviour
started occuring out of nowhere without me doing anything no config
change etc,

I suspect a few things, one been mysql since the 2 busiest sites are
forums and the other been the hard drive as I noticed weird pauses
when doing tasks every now and then even when the server is low load.

What also makes me suspect the hd is that cpu usage is often high on
system especially when its clogging up, system is often higher then
user.

Here is a typical top output.

last pid: 28474;  load averages:  0.47,  0.33,  0.36    up 0+21:15:28  06:57:22
144 processes: 1 running, 143 sleeping
CPU states:  0.8% user,  0.0% nice,  0.4% system,  0.4% interrupt, 98.4% idle
Mem: 486M Active, 788M Inact, 241M Wired, 91M Cache, 112M Buf, 270M Free
Swap: 2048M Total, 52M Used, 1996M Free, 2% Inuse

  PID USERNAME   PRI NICE   SIZE    RES STATE    TIME   WCPU    CPU COMMAND
93241 apache      20    0   110M 69636K lockf    0:15  2.05%  2.05% httpd
56049 apache      20    0   110M 72544K lockf    0:19  1.37%  1.37% httpd
37231 apache      20    0   110M 68764K lockf    0:19  0.93%  0.93% httpd
92245 root         4   15 11220K 10336K kqread   0:28  0.49%  0.49% perl5.8.8
51077 apache      20    0   110M 56552K lockf    0:04  0.49%  0.49% httpd
56293 apache      20    0   110M 73752K lockf    0:22  0.34%  0.34% httpd
 4289 root        96    0 30040K 22624K select   1:09  0.29%  0.29% perl
44432 apache      20    0   110M 73012K lockf    0:22  0.29%  0.29% httpd
49998 apache      20    0   114M 76828K lockf    0:22  0.24%  0.24% httpd
49797 apache      20    0   110M 71412K lockf    0:21  0.15%  0.15% httpd
92511 apache      20    0   139M 97804K lockf    0:10  0.10%  0.10% httpd
25367 mysql       20   -5   261M 98408K kserel  10:13  0.05%  0.05% mysqld
31309 apache      20    0   109M 71664K lockf    0:16  0.05%  0.05% httpd
64100 apache      20    0   139M   112M lockf    1:46  0.00%  0.00% httpd
10469 root        96    0 32664K 20860K select   0:39  0.00%  0.00% perl
45354 apache      20    0   110M 76968K lockf    0:23  0.00%  0.00% httpd

The only reason swap is in use is because it never flushed fully from
the last buildup of processes.  My httpd.conf setting for maxclients
etc. is as below.

Timeout 15
KeepAlive Off
MaxKeepAliveRequests 150
KeepAliveTimeout 5
MinSpareServers 8
MaxSpareServers 16
StartServers 8
MaxClients 100
MaxRequestsPerChild 5000

They are not permanent settings it previously has also had much higher
Timeout of 60 and Maxrequestsperchild I have tried lower on 50.  It is
a single cpu machine.

Chris

---------------------------------------------------------------------
The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
   "   from the digest: users-digest-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org


Mime
View raw message