Hello Rainer Jung,
Thanks for the response.
I can see there are many threads like below which are in sleep state.
I have taken thread dumps 4 or 5 times and can see this threads in all the =
dumps.
For the instance i am pasting some of similar kind of threads from the thre=
ad dump.
I have to show this to developers but before that i have to be very much su=
re there is no issue from apache web server or tomcat application server en=
d.
"TP-Processor1" daemon prio=3D5 tid=3D0x000a8538 nid=3D0x24 waiting on cond=
ition [9747e000..9747fc30]
=A0=A0=A0 at java.lang.Thread.sleep(Native Method)
=A0=A0=A0 at com.ericsson.eai.b2b.MUXServlet.doPost(MUXServlet.java:506)
=A0=A0=A0 at javax.servlet.http.HttpServlet.service(HttpServlet.java:709)
=A0=A0=A0 at javax.servlet.http.HttpServlet.service(HttpServlet.java:802)
=A0=A0=A0 at org.apache.catalina.core.ApplicationFilterChain.internalDoFilt=
er(ApplicationFilterChain.java:252)
=A0=A0=A0 at org.apache.catalina.core.ApplicationFilterChain.doFilter(Appli=
cationFilterChain.java:173)
=A0=A0=A0 at org.apache.catalina.core.StandardWrapperValve.invoke(StandardW=
rapperValve.java:213)
=A0=A0=A0 at org.apache.catalina.core.StandardContextValve.invoke(StandardC=
ontextValve.java:178)
=A0=A0=A0 at org.apache.catalina.core.StandardHostValve.invoke(StandardHost=
Valve.java:126)
=A0=A0=A0 at org.apache.catalina.valves.ErrorReportValve.invoke(ErrorReport=
Valve.java:105)
=A0=A0=A0 at org.apache.catalina.core.StandardEngineValve.invoke(StandardEn=
gineValve.java:107)
=A0=A0=A0 at org.apache.catalina.connector.CoyoteAdapter.service(CoyoteAdap=
ter.java:148)
=A0=A0=A0 at org.apache.jk.server.JkCoyoteHandler.invoke(JkCoyoteHandler.ja=
va:307)
=A0=A0=A0 at org.apache.jk.common.HandlerRequest.invoke(HandlerRequest.java=
:385)
=A0=A0=A0 at org.apache.jk.common.ChannelSocket.invoke(ChannelSocket.java:7=
48)
=A0=A0=A0 at org.apache.jk.common.ChannelSocket.processConnection(ChannelSo=
cket.java:678)
=A0=A0=A0 at org.apache.jk.common.SocketConnection.runIt(ChannelSocket.java=
:871)
=A0=A0=A0 at org.apache.tomcat.util.threads.ThreadPool$ControlRunnable.run(=
ThreadPool.java:684)
=A0=A0=A0 at java.lang.Thread.run(Thread.java:534)
"TP-Processor2" daemon prio=3D5 tid=3D0x000a90c0 nid=3D0x25 waiting on cond=
ition [9737e000..9737fc30]
=A0=A0=A0 at java.lang.Thread.sleep(Native Method)
=A0=A0=A0 at com.ericsson.eai.b2b.MUXServlet.doPost(MUXServlet.java:506)
=A0=A0=A0 at javax.servlet.http.HttpServlet.service(HttpServlet.java:709)
=A0=A0=A0 at javax.servlet.http.HttpServlet.service(HttpServlet.java:802)
=A0=A0=A0 at org.apache.catalina.core.ApplicationFilterChain.internalDoFilt=
er(ApplicationFilterChain.java:252)
=A0=A0=A0 at org.apache.catalina.core.ApplicationFilterChain.doFilter(Appli=
cationFilterChain.java:173)
=A0=A0=A0 at org.apache.catalina.core.StandardWrapperValve.invoke(StandardW=
rapperValve.java:213)
=A0=A0=A0 at org.apache.catalina.core.StandardContextValve.invoke(StandardC=
ontextValve.java:178)
=A0=A0=A0 at org.apache.catalina.core.StandardHostValve.invoke(StandardHost=
Valve.java:126)
=A0=A0=A0 at org.apache.catalina.valves.ErrorReportValve.invoke(ErrorReport=
Valve.java:105)
=A0=A0=A0 at org.apache.catalina.core.StandardEngineValve.invoke(StandardEn=
gineValve.java:107)
=A0=A0=A0 at org.apache.catalina.connector.CoyoteAdapter.service(CoyoteAdap=
ter.java:148)
=A0=A0=A0 at org.apache.jk.server.JkCoyoteHandler.invoke(JkCoyoteHandler.ja=
va:307)
=A0=A0=A0 at org.apache.jk.common.HandlerRequest.invoke(HandlerRequest.java=
:385)
=A0=A0=A0 at org.apache.jk.common.ChannelSocket.invoke(ChannelSocket.java:7=
48)
=A0=A0=A0 at org.apache.jk.common.ChannelSocket.processConnection(ChannelSo=
cket.java:678)
=A0=A0=A0 at org.apache.jk.common.SocketConnection.runIt(ChannelSocket.java=
:871)
=A0=A0=A0 at org.apache.tomcat.util.threads.ThreadPool$ControlRunnable.run(=
ThreadPool.java:684)
=A0=A0=A0 at java.lang.Thread.run(Thread.java:534)
--- On Tue, 1/11/11, Rainer Jung <rainer.jung@kippdata.de> wrote:
From: Rainer Jung <rainer.jung@kippdata.de>
Subject: Re: Apache Process going high even after restart of backend server=
tomcat
To: "Tomcat Users List" <users@tomcat.apache.org>
Date: Tuesday, January 11, 2011, 10:51 PM
On 10.01.2011 17:00, Amol Puglia wrote:
> Hello Team,
>=20
> I am facing very weird problem of apache porcess going high upto 259 and =
increasing again even after restart of backend server ie. tomcat.
>=20
> Apache version :- 2.0.59
>=20
> Tomcat Version :- 5.5.9
>=20
> Apache/Tomcat connectivity :- mod_jk
>=20
> when i have checked mod_jk.log file i am seeing this messages in the logs=
.
>=20
> [Mon Jan 10 08:14:52 2011] worker1 erib2bdev.ericsson.net 0.001122
> [Mon Jan 10 08:14:55 2011] worker1 erib2bdev.ericsson.net 0.000732
> [Mon Jan 10 08:14:57 2011] worker1 erib2bdev.ericsson.net 0.000740
> [Mon Jan 10 08:15:00 2011] worker1 erib2bdev.ericsson.net 0.000846
> [Mon Jan 10 08:15:02 2011] worker1 erib2bdev.ericsson.net 0.000952
As I already wrote on the httpd users list:
The lines come from the JkRequestLogFormat configuration setting that you m=
ust have added somewhere (default: no such setting, no lines). The configur=
ed format will tell you, what the entry means. Likely it is either the name=
of the virtual host, or the value of the host header the client sent. Clea=
rtext: this is requests being processed by your Apache server. The client s=
ends these requests to "erib2bdev.ericsson.net" and they are received by yo=
ur server. Or you configured your server to carry the name "erib2bdev.erics=
son.net".
For details see:
http://tomcat.apache.org/connectors-doc/reference/apache.html
and look for "JkRequestLogFormat" in that page.
The use of that directive is no longer recommended, because it is easier to=
include request log information into the usual access log via mod_log_conf=
ig. Look for "mod_log_config" in the same page.
> I am not sure what does it mean when the name "erib2bdev.ericsson.net" is=
not refereed anywhere in my apache and tomcat configuration files and not =
even in workers.properties file. Can anyone help me torubleshoot this issue=
?
Furthermore this might not have to do with your issue, namely that you thin=
k the number of Apache processes is to big. From the other thread you poste=
d containing the Java thread Dump excerpt we think that the reason for the =
growing number of processes is, that your MUXServlet doesn't handle the req=
uest and instead enters sleep. To make this assumption a fact, you would ne=
ed to tell us, whether there are many threads in the same dump, that have t=
he same stack (sleep in MUXServlet), more precisely how many and what the n=
ames of the threads are. If you find let's say 200 of those, then you are c=
lose to an explanation.
Finally you have to do more than one dump to understand, whether the sleep(=
) was only lasting for a short time, or it is happening all the time. Then =
you have to start talking to the developers of the MUXServlet showing them =
those dumps.
Regards,
Rainer
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org
=0A=0A=0A
|