Thanks for your reply!
Keep alive is disabled. The slower response times are seldom, nearly 0,1
% of all the requests. The tests were done with a concurrency of 1.
I think, as Chris replied, it 's another process of the OS.
the server.xml:
<Server port="8005" shutdown="SHUTDOWN">
<Listener className="org.apache.catalina.mbeans.ServerLifecycleListener" />
<Listener
className="org.apache.catalina.mbeans.GlobalResourcesLifecycleListener" />
<Listener
className="org.apache.catalina.storeconfig.StoreConfigLifecycleListener" />
<Listener className="org.apache.catalina.mbeans.JMXAdaptorLifecycleListener"
namingPort="8890" port="8891" host="servername />
<GlobalNamingResources>
<Resource name="UserDatabase" auth="Container"
type="org.apache.catalina.UserDatabase" description="User database that
can be updated and saved"
factory="org.apache.catalina.users.MemoryUserDatabaseFactory"
pathname="conf/tomcat-users.xml" />
</GlobalNamingResources>
<Service name="Catalina">
<Connector port="80" URIEncoding="UTF-8" maxHttpHeaderSize="8192"
maxThreads="3000" minSpareThreads="3000" maxSpareThreads="3000"
enableLookups="false" acceptCount="25"
connectionTimeout="10000" disableUploadTimeout="true" server="server"
maxKeepAliveRequests="1"
compression="on" compressionMinSize="2048"
noCompressionUserAgents="gozilla, traviata"
compressableMimeType="text/html,text/xml,text/css" />
<Engine name="Catalina" defaultHost="localhost">
<Realm className="org.apache.catalina.realm.UserDatabaseRealm"
resourceName="UserDatabase" />
<Host name="localhost" appBase="webapps" unpackWARs="true"
autoDeploy="true" xmlValidation="false" xmlNamespaceAware="false">
<Valve className="org.apache.catalina.valves.AccessLogValve"
directory="logs" prefix="localhost_access_log." suffix=".txt"
pattern="%h %l %u %t
"%r" %s %b %D" resolveHosts="false" />
</Host>
</Engine>
</Service>
</Server>
Regards
Peter
Am 03.08.2010 22:39, schrieb André Warnier:
Peter Buning wrote:
Hello,
we tested our tomcat 5.5 server (Debian system) with the apache
benchmark tool. We send requests for a small png-file (300 bytes).
Nearly all requests were done in 0 or 1 milliseconds. But there were
a few request which needed more than 10 ms and the slowest request
needed 106 ms.
the 106 ms is logged in the access log file. We used the %D pattern
for the request time in millis.
It seems not to be a network problem. We can see the same behavior
with a remote and a localhost test. Nevertheless, does anybody know,
when tomact exactly starts and stop the time for logging the request?
We did A similar test with the YourKit Java Profiler, but can't find
a reason for the slow requests, too. The time for "all threads" and
for "java.lang.Thread.run" was less than the time for the longest
request. The Garbage Collector was not active while testing.
So, we want to know, why there are a few request that took so much
more time than the rest of the requests. Has anybody an idea where
the time can be consumed?
Are the requests from the benchmark tool made with "keep-alive" ?
What is the keepAlive setting for the Tomcat Connector processing
these requests ?
Depending on your configuration and how requests are being made, if
keepAlive is set to a high value, you may just have many Tomcat
threads waiting for a follow-up request (*) on each connection, before
closing the connection and becoming available again to process more
requests.
(*) which may never come..
In your next message, paste the contents of your Tomcat's server.xml
file, all comments and private data removed.
Also tell how the requests are being issued (how fast, how many,
with/without keepalive, etc..)
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
For additional commands, e-mail: users-h...@tomcat.apache.org
