Hi,
Just as a kind of slightly-different sanity check, I did some testing on an 
Apache 2.2.31 instance that I built awhile ago on one of my test CENTOS 
machines:
Server version: Apache/2.2.31 (Unix)Server built:   Oct  7 2015 
11:32:17Server's Module Magic Number: 20051115:40Server loaded:  APR 1.5.2, 
APR-Util 1.5.4Compiled using: APR 1.5.2, APR-Util 1.5.4Architecture:   
64-bitServer MPM:     Prefork  threaded:     no    forked:     yes (variable 
process count)Server compiled with.... -D APACHE_MPM_DIR="server/mpm/prefork" 
-D APR_HAS_SENDFILE -D APR_HAS_MMAP -D APR_HAVE_IPV6 (IPv4-mapped addresses 
enabled) -D APR_USE_SYSVSEM_SERIALIZE -D APR_USE_PTHREAD_SERIALIZE -D 
SINGLE_LISTEN_UNSERIALIZED_ACCEPT -D APR_HAS_OTHER_CHILD -D 
AP_HAVE_RELIABLE_PIPED_LOGS -D DYNAMIC_MODULE_LIMIT=128 -D 
HTTPD_ROOT="/apps/httpd" -D SUEXEC_BIN="/apps/httpd/bin/suexec" -D 
DEFAULT_PIDLOG="logs/httpd.pid" -D 
DEFAULT_SCOREBOARD="logs/apache_runtime_status" -D 
DEFAULT_LOCKFILE="logs/accept.lock" -D DEFAULT_ERRORLOG="logs/error_log" -D 
AP_TYPES_CONFIG_FILE="conf/mime.types" -D SERVER_CONFIG_FILE="conf/httpd.conf"
Then, I ran essentially the same load test that I had been running earlier, 
and, I get similar errors, e.g., connection refused.
This is with the as-built Apache configuration settings (i.e., I haven't tried 
tweaking them).
This is kind of leading me to think that it may the Apache, built with the 
above MPM model, may just not be able to sustain this kind of load error-free? 
Thanks,Jim
      From: o haya <oh...@yahoo.com.INVALID>
 To: "users@httpd.apache.org" <users@httpd.apache.org> 
Cc: O. Haya <oh...@yahoo.com> 
 Sent: Monday, November 16, 2015 3:44 PM
 Subject: Re: [users@httpd] Getting "connection refused" Apache 2.4
   
Hi Daniel,
Unfortunately, the Apache I'm working with is part of a product, so I won't be 
able to change to a different MPM type.
I've been trying what you suggested (increasing MinSpareServers) but I still 
get tons of connections refused, even when I can confirm that, for example, I 
get 500+ httpd instances at startup (e.g., ps -ef shows 500 instances).  Is it 
possible that something else (e.g., the Redhat OS itself) is limiting the 
number of connections?
Thanks,Jim

 

     From: Daniel <dferra...@gmail.com>
 To: users@httpd.apache.org; o haya <oh...@yahoo.com> 
 Sent: Monday, November 16, 2015 3:07 PM
 Subject: Re: [users@httpd] Getting "connection refused" Apache 2.4
   
migrate to worker or event...if you have to stick with prefork mpm, increase 
the number of MinSpareServers to a very high number, prefork is very slow 
spawning new children since it is not threaded, so better spawn most when 
apache starts.


El lun., 16 nov. 2015 a las 20:58, o haya (<oh...@yahoo.com.invalid>) escribió:

Hi,

I am trying to do some load testing of Apache 2.4 on Redhat.

Configuration displayed when I run "httpd -V" shows:

Server MPM: prefork
threaded: no
forked: yes (variable process count)

I am using Jmeter to test, and when I increase the number of simultaneous 
threads to > 200, I am seeing increasing number of errors, with a lot of 
"connection refused" errors, e.g.:

org.apache.http.conn.HttpHostConnectException: Connection to 
http://10.0.3.10:80 refused
        at 
org.apache.http.impl.conn.DefaultClientConnectionOperator.openConnection(DefaultClientConnectionOperator.java:190)
        at 
org.apache.http.impl.conn.ManagedClientConnectionImpl.open(ManagedClientConnectionImpl.java:294)
        at 
org.apache.jmeter.protocol.http.sampler.MeasuringConnectionManager$MeasuredConnection.open(MeasuringConnectionManager.java:107)
        at 
org.apache.http.impl.client.DefaultRequestDirector.tryConnect(DefaultRequestDirector.java:643)
        at 
org.apache.http.impl.client.DefaultRequestDirector.execute(DefaultRequestDirector.java:479)
        at 
org.apache.http.impl.client.AbstractHttpClient.execute(AbstractHttpClient.java:906)
        at 
org.apache.http.impl.client.AbstractHttpClient.execute(AbstractHttpClient.java:805)
        at 
org.apache.jmeter.protocol.http.sampler.HTTPHC4Impl.executeRequest(HTTPHC4Impl.java:517)
        at 
org.apache.jmeter.protocol.http.sampler.HTTPHC4Impl.sample(HTTPHC4Impl.java:331)
        at 
org.apache.jmeter.protocol.http.sampler.HTTPSamplerProxy.sample(HTTPSamplerProxy.java:74)
        at 
org.apache.jmeter.protocol.http.sampler.HTTPSamplerBase.sample(HTTPSamplerBase.java:1146)
        at 
org.apache.jmeter.protocol.http.sampler.HTTPSamplerBase.sample(HTTPSamplerBase.java:1135)
        at 
org.apache.jmeter.threads.JMeterThread.process_sampler(JMeterThread.java:434)
        at org.apache.jmeter.threads.JMeterThread.run(JMeterThread.java:261)
        at java.lang.Thread.run(Unknown Source)
Caused by: java.net.ConnectException: Connection timed out: connect
        at java.net.DualStackPlainSocketImpl.waitForConnect(Native Method)
        at java.net.DualStackPlainSocketImpl.socketConnect(Unknown Source)
        at java.net.AbstractPlainSocketImpl.doConnect(Unknown Source)
        at java.net.AbstractPlainSocketImpl.connectToAddress(Unknown Source)
        at java.net.AbstractPlainSocketImpl.connect(Unknown Source)
        at java.net.PlainSocketImpl.connect(Unknown Source)
        at java.net.SocksSocketImpl.connect(Unknown Source)
        at java.net.Socket.connect(Unknown Source)
        at 
org.apache.http.conn.scheme.PlainSocketFactory.connectSocket(PlainSocketFactory.java:127)
        at 
org.apache.http.impl.conn.DefaultClientConnectionOperator.openConnection(DefaultClientConnectionOperator.java:180)
        ... 14 more


I've been trying to re-configure the Apache to avoid these errors, e.g., 
increasing "MaxClients", and I've also increased the amount of memory for the 
machine (to 4GB) but so far, nothing I have done has improved the situation.

Is there something else that I can do to try to get the Apache so that it can 
accept the larger number of connections?

Thanks,
Jim



---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@httpd.apache.org
For additional commands, e-mail: users-h...@httpd.apache.org




   

  

Reply via email to