You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@tomcat.apache.org by me...@assetmanagement.natexis.fr on 2004/02/20 10:25:27 UTC
Réf. : Re: maxProcessors = max ammout ofconcurrent connections?
As I know, the maxProcessor value must match
the MaxClients value in apache, because each apache
process must connect to a processor in Tomcat.
_______________________________________________
NATEXIS ASSET MANAGEMENT
Meissa SAKHO
01 58 19 45 71. . . . . . . . . . . . (84571)
meissa.sakho@assetmanagement.natexis.fr
"Christian Hufgard" <ch...@gmx.de>
20/02/2004 10:03
Veuillez répondre à "Tomcat Users List"
Pour : "Tomcat Users List" <to...@jakarta.apache.org>
cc :
Objet : Re: maxProcessors = max ammout of concurrent connections?
> I have several machines running tomcat now, but I like to get it tuned
for
> higher load. How can I find out the maximum processors I shoudl accept
?
> Just putting in a high number of course results in very slow processing
of
> request, resulting in people refreshing, results in even more processes,
> results in even slower replies, results in more refreshers :)
Well, think you should not set the max processors to high, that the
machine
takes longer time to handle the requests. You should define your maximum
response time and the figure out, how many request you can process at once
in
this time.
> Any idea how to 'tune' the Maximum connectors and the acceptcount. Is
it
> really needed to have an acceptcount ?
If the acceptCount is zero, the user will get an message, saying him that
the server cannot be reached. If he is in the accepted queue, he will be
shown
that the server has been contacted and is about to process his request.
Christian
> Is there any document 'out in the wild' describing how to know which
> values you should use ?
>
> I'm just getting to know tomcat but I find it hard to find the right
> documentation. (many books describe the options in a complete
> different/confising way).
>
> I have several machines running tomcat now, but I like to get it tuned
for
> higher load. How can I find out the maximum processors I shoudl accept
?
> Just putting in a high number of course results in very slow processing
of
> request, resulting in people refreshing, results in even more processes,
> results in even slower replies, results in more refreshers :)
>
> Any idea how to 'tune' the Maximum connectors and the acceptcount. Is
it
> really needed to have an acceptcount ?
>
> Pepijn
>
> On Fri, 20 Feb 2004, Christian Hufgard wrote:
>
> > > Why put minProcessors higher then the maximum allowed ?
> >
> > Think this does make no sense and should result in either a constant
> ammount
> > of processors (the min value, since it cannot fall lower), an error
> message
> > at startup or some other, unexpected behaviour.
> > Now I set the minProcessors to 30, maxProcessor to 10.
> > What happens is: No error messages in any log.
> > 8 connections after startup of my prog.
> > 9th connection after two seconds
> > 10th connections after nine seconds.
> > Then with each second a new connection is accepted. (Stopped the prog
at
> 20
> > connections)
> >
> > After restarting 18 connections are established directly, two more
after
> > some seconds.
> >
> > 50 seconds after starting the program:
> > Connections are interrupted. Logfile says:
> > INFO: All threads are busy, waiting. Please increase maxThreads or
check
> the
> > servlet status10 10.
> >
> >
> >
> >
> > > I'm having trouble finding the exact meaning of some of the
parameters
> > > myself... but this might be part of a solution for your problem:
> > >
> > > Why put minProcessors higher then the maximum allowed ?
> > >
> > > Pepijn Palmans
> > > Managing Director
> > >
> > > Kangaroot Linux Solutions
> > > Grote Steenweg 91
> > > 2600 Antwerp, Belgium
> > > Tel: +32 3/286.17.17
> > > Fax: +32 3/281.23.49
> > >
> > >
> > > On Fri, 20 Feb 2004, Christian Hufgard wrote:
> > >
> > > > Hi folks,
> > > >
> > > > did not found this question answered, neither in the faqs, nor in
> the
> > > > archive of this mailing list, nor in bugzilla.
> > > >
> > > > Like in the subject described, my question is, wether the
attribute
> > > > maxProcessors in the Connector tag sets the maximum ammount of
> > > concurrent connections
> > > > the CoyeteConnector handles.
> > > >
> > > > I tested this on Tomcat up to 4.1.24 with jdk 1.4.2 on debian with
> > > kernel
> > > > 2.4.19, since this is our productive environment. If i set up the
> > > connector in
> > > > server.xml with the following parameters:
> > > >
> > > > <!-- Define a non-SSL Coyote HTTP/1.1 Connector on port 8080
-->
> > > > <Connector
className="org.apache.coyote.tomcat4.CoyoteConnector"
> > > > port="8080" minProcessors="20" maxProcessors="10"
> > > > enableLookups="true" redirectPort="8443"
> > > > acceptCount="5" debug="0" connectionTimeout="20000"
> > > > useURIValidationHack="false"
> disableUploadTimeout="true"
> > > />
> > > >
> > > > I find a pretty strange behaviour (if maxProcessors should be max
> > > > Connections):
> > > > My test programm (jdk 1.4) tries to open up 80 connections to the
> tomcat
> > > and
> > > > send some continueus requests to a webapps that create some load
on
> the
> > > > server. The requests use HTTP/1.1 to keep the connection opened
and
> > > request a
> > > > page each half second with variing parameters..
> > > > I would expect, that 15 connections would be opened. 10 active + 5
> on
> > > hold.
> > > > What I see is, that normally 18 connections are accepted upon
> startup of
> > > my
> > > > programm. After 8 seconds another two connections are established,
> 12
> > > seconds
> > > > later, again two new connections... This leads up to about 30-40
> > > > connections.
> > > >
> > > > So, finally I precise my question a bit: What exactly does
> maxProcessors
> > > > mean? If it does not mean max concurrent connections, how many
> > > connections can
> > > > one processor handle?
> > > >
> > > > Hope I could describe my problem fairly enough :)
> > > >
> > > > Greets,
> > > >
> > > > Christian
> > > >
> > > > --
> > > > GMX ProMail (250 MB Mailbox, 50 FreeSMS, Virenschutz, 2,99
> EUR/Monat...)
> > > > jetzt 3 Monate GRATIS + 3x DER SPIEGEL +++
> http://www.gmx.net/derspiegel
> > > +++
> > > >
> > > >
> > > >
> ---------------------------------------------------------------------
> > > > To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
> > > > For additional commands, e-mail:
tomcat-user-help@jakarta.apache.org
> > > >
> > >
> > >
---------------------------------------------------------------------
> > > To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
> > > For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
> > >
> >
> > --
> > GMX ProMail (250 MB Mailbox, 50 FreeSMS, Virenschutz, 2,99
EUR/Monat...)
> > jetzt 3 Monate GRATIS + 3x DER SPIEGEL +++ http://www.gmx.net/derspiegel
> +++
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
> > For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
>
--
GMX ProMail (250 MB Mailbox, 50 FreeSMS, Virenschutz, 2,99 EUR/Monat...)
jetzt 3 Monate GRATIS + 3x DER SPIEGEL +++ http://www.gmx.net/derspiegel +++
---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
L'integrite de ce message n'etant pas assuree sur internet, Natexis
Banques Populaires ne peut etre tenu responsable de
son contenu. Toute utilisation ou diffusion non autorisee est
interdite. Si vous n'etes pas destinataire de ce message, merci de le
detruire et d'avertir l'expediteur.
The integrity of this message cannot be guaranteed
on the Internet. Natexis Banques Populaires can not therefore be
considered responsible for the contents.Any unauthorized use or dissemination is prohibited.
If you are not the intended recipient of this message, then please delete it and
notify the sender.
Re: Réf. : Re: maxProcessors = max ammout of concurrent connections?
Posted by Christian Hufgard <ch...@gmx.de>.
We serve all content via tomcat, so here is no apache involved.
The non-static content would be only a fiew images and some css, while
nearly everyting else in generated dynamically or has to be proteced via our
framework.
Christian
> That is not true. This can be only true if you have only one instance of
> tomcat, and ALL (even static!!!) content is serverd by tomcat.
>
> Under normal setup you let your static content be served by apache, and
> only use tomcat when needed. So you might wanna have more apache
> processes.
>
> Pepijn
>
> On Fri, 20 Feb 2004 meissa.Sakho@assetmanagement.natexis.fr wrote:
>
> >
> > As I know, the maxProcessor value must match
> > the MaxClients value in apache, because each apache
> > process must connect to a processor in Tomcat.
> > _______________________________________________
> > NATEXIS ASSET MANAGEMENT
> > Meissa SAKHO
> > 01 58 19 45 71. . . . . . . . . . . . (84571)
> > meissa.sakho@assetmanagement.natexis.fr
> >
> >
> >
> >
> > "Christian Hufgard" <ch...@gmx.de>
> > 20/02/2004 10:03
> > Veuillez répondre à "Tomcat Users List"
> >
> >
> > Pour : "Tomcat Users List" <to...@jakarta.apache.org>
> > cc :
> > Objet : Re: maxProcessors = max ammout of concurrent connection
> s?
> >
> >
> > > I have several machines running tomcat now, but I like to get it tuned
> > for
> > > higher load. How can I find out the maximum processors I shoudl
> accept
> > ?
> > > Just putting in a high number of course results in very slow
> processing
> > of
> > > request, resulting in people refreshing, results in even more
> processes
> ,
> > > results in even slower replies, results in more refreshers :)
> >
> > Well, think you should not set the max processors to high, that the
> > machine
> > takes longer time to handle the requests. You should define your maximum
> > response time and the figure out, how many request you can process at
> onc
> e
> > in
> > this time.
> >
> >
> > > Any idea how to 'tune' the Maximum connectors and the acceptcount. Is
> > it
> > > really needed to have an acceptcount ?
> > If the acceptCount is zero, the user will get an message, saying him
> that
> > the server cannot be reached. If he is in the accepted queue, he will be
> > shown
> > that the server has been contacted and is about to process his request.
> >
> > Christian
> >
> >
> >
> >
> > > Is there any document 'out in the wild' describing how to know which
> > > values you should use ?
> > >
> > > I'm just getting to know tomcat but I find it hard to find the right
> > > documentation. (many books describe the options in a complete
> > > different/confising way).
> > >
> > > I have several machines running tomcat now, but I like to get it tuned
> > for
> > > higher load. How can I find out the maximum processors I shoudl
> accept
> > ?
> > > Just putting in a high number of course results in very slow
> processing
> > of
> > > request, resulting in people refreshing, results in even more
> processes
> ,
> > > results in even slower replies, results in more refreshers :)
> > >
> > > Any idea how to 'tune' the Maximum connectors and the acceptcount. Is
> > it
> > > really needed to have an acceptcount ?
> > >
> > > Pepijn
> > >
> > > On Fri, 20 Feb 2004, Christian Hufgard wrote:
> > >
> > > > > Why put minProcessors higher then the maximum allowed ?
> > > >
> > > > Think this does make no sense and should result in either a constant
> > > ammount
> > > > of processors (the min value, since it cannot fall lower), an error
> > > message
> > > > at startup or some other, unexpected behaviour.
> > > > Now I set the minProcessors to 30, maxProcessor to 10.
> > > > What happens is: No error messages in any log.
> > > > 8 connections after startup of my prog.
> > > > 9th connection after two seconds
> > > > 10th connections after nine seconds.
> > > > Then with each second a new connection is accepted. (Stopped the
> prog
> > at
> > > 20
> > > > connections)
> > > >
> > > > After restarting 18 connections are established directly, two more
> > after
> > > > some seconds.
> > > >
> > > > 50 seconds after starting the program:
> > > > Connections are interrupted. Logfile says:
> > > > INFO: All threads are busy, waiting. Please increase maxThreads or
> > check
> > > the
> > > > servlet status10 10.
> > > >
> > > >
> > > >
> > > >
> > > > > I'm having trouble finding the exact meaning of some of the
> > parameters
> > > > > myself... but this might be part of a solution for your problem:
> > > > >
> > > > > Why put minProcessors higher then the maximum allowed ?
> > > > >
> > > > > Pepijn Palmans
> > > > > Managing Director
> > > > >
> > > > > Kangaroot Linux Solutions
> > > > > Grote Steenweg 91
> > > > > 2600 Antwerp, Belgium
> > > > > Tel: +32 3/286.17.17
> > > > > Fax: +32 3/281.23.49
> > > > >
> > > > >
> > > > > On Fri, 20 Feb 2004, Christian Hufgard wrote:
> > > > >
> > > > > > Hi folks,
> > > > > >
> > > > > > did not found this question answered, neither in the faqs, nor
> in
> > > the
> > > > > > archive of this mailing list, nor in bugzilla.
> > > > > >
> > > > > > Like in the subject described, my question is, wether the
> > attribute
> > > > > > maxProcessors in the Connector tag sets the maximum ammount of
> > > > > concurrent connections
> > > > > > the CoyeteConnector handles.
> > > > > >
> > > > > > I tested this on Tomcat up to 4.1.24 with jdk 1.4.2 on debian
> wit
> h
> > > > > kernel
> > > > > > 2.4.19, since this is our productive environment. If i set up
> the
> > > > > connector in
> > > > > > server.xml with the following parameters:
> > > > > >
> > > > > > <!-- Define a non-SSL Coyote HTTP/1.1 Connector on port 8080
> > -->
> > > > > > <Connector
> > className="org.apache.coyote.tomcat4.CoyoteConnector"
> > > > > > port="8080" minProcessors="20" maxProcessors
> ="10"
> > > > > > enableLookups="true" redirectPort="8443"
> > > > > > acceptCount="5" debug="0" connectionTimeout=
> "20000"
> > > > > > useURIValidationHack="false"
> > > disableUploadTimeout="true"
> > > > > />
> > > > > >
> > > > > > I find a pretty strange behaviour (if maxProcessors should be
> max
> > > > > > Connections):
> > > > > > My test programm (jdk 1.4) tries to open up 80 connections to
> the
> > > tomcat
> > > > > and
> > > > > > send some continueus requests to a webapps that create some load
> > on
> > > the
> > > > > > server. The requests use HTTP/1.1 to keep the connection opened
> > and
> > > > > request a
> > > > > > page each half second with variing parameters..
> > > > > > I would expect, that 15 connections would be opened. 10 active +
> 5
> > > on
> > > > > hold.
> > > > > > What I see is, that normally 18 connections are accepted upon
> > > startup of
> > > > > my
> > > > > > programm. After 8 seconds another two connections are
> established
> ,
> > > 12
> > > > > seconds
> > > > > > later, again two new connections... This leads up to about 30-40
> > > > > > connections.
> > > > > >
> > > > > > So, finally I precise my question a bit: What exactly does
> > > maxProcessors
> > > > > > mean? If it does not mean max concurrent connections, how many
> > > > > connections can
> > > > > > one processor handle?
> > > > > >
> > > > > > Hope I could describe my problem fairly enough :)
> > > > > >
> > > > > > Greets,
> > > > > >
> > > > > > Christian
> > > > > >
> > > > > > --
> > > > > > GMX ProMail (250 MB Mailbox, 50 FreeSMS, Virenschutz, 2,99
> > > EUR/Monat...)
> > > > > > jetzt 3 Monate GRATIS + 3x DER SPIEGEL +++
> > > http://www.gmx.net/derspiegel
> > > > > +++
> > > > > >
> > > > > >
> > > > > >
> > > ---------------------------------------------------------------------
> > > > > > To unsubscribe, e-mail:
> tomcat-user-unsubscribe@jakarta.apache.or
> g
> > > > > > For additional commands, e-mail:
> > tomcat-user-help@jakarta.apache.org
> > > > > >
> > > > >
> > > > >
> > ---------------------------------------------------------------------
> > > > > To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
> > > > > For additional commands, e-mail:
> tomcat-user-help@jakarta.apache.or
> g
> > > > >
> > > >
> > > > --
> > > > GMX ProMail (250 MB Mailbox, 50 FreeSMS, Virenschutz, 2,99
> > EUR/Monat...)
> > > > jetzt 3 Monate GRATIS + 3x DER SPIEGEL +++
> http://www.gmx.net/derspie
> gel
> > > +++
> > > >
> > > >
> > > >
> ---------------------------------------------------------------------
> > > > To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
> > > > For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
> > > >
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
> > > For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
> > >
> >
> > --
> > GMX ProMail (250 MB Mailbox, 50 FreeSMS, Virenschutz, 2,99 EUR/Monat...)
> > jetzt 3 Monate GRATIS + 3x DER SPIEGEL +++ http://www.gmx.net/derspiegel
> +++
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
> > For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
> >
> >
> >
> >
> >
> > L'integrite de ce message n'etant pas assuree sur internet, Natexis
> > Banques Populaires ne peut etre tenu responsable de
> > son contenu. Toute utilisation ou diffusion non autorisee est
> > interdite. Si vous n'etes pas destinataire de ce message, merci de le
> > detruire et d'avertir l'expediteur.
> >
> > The integrity of this message cannot be guaranteed
> > on the Internet. Natexis Banques Populaires can not therefore be
> > considered responsible for the contents.Any unauthorized use or
> dissemina
> tion is prohibited.
> > If you are not the intended recipient of this message, then please
> delete
> it and
> > notify the sender.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
>
--
GMX ProMail (250 MB Mailbox, 50 FreeSMS, Virenschutz, 2,99 EUR/Monat...)
jetzt 3 Monate GRATIS + 3x DER SPIEGEL +++ http://www.gmx.net/derspiegel +++
---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
Re: Réf. : Re: maxProcessors = max ammout of concurrent connections?
Posted by Antonio Fiol Bonnín <fi...@terra.es>.
As far as "connections" does not mean "active connections", what
"meissa.Sakho" wrote is strictly true.
Each Apache process connects to Tomcat when (if) it needs to. Call that
a "connection".
Apache processes do not disconnect from Tomcat. Call those "idle
connections".
idle connections <= connections <= number of Apache processes <= MaxClients
Of course some of the connections are used at a certain moment.
active connections <= connections
But at any moment in time, an Apache process that is not connected to
Tomcat MAY receive a request, and start a connection to Tomcat. So, it
is possible (and it actually happens most of the time) that even if your
Apache is serving static content, all of the Apache processes are
actually connected to Tomcat.
This means that connections <= MaxClients, with the strict equality
being possible.
If you have more than one instance of Tomcat, the same thing happens.
An Apache process that is connected to one of your Tomcat servers MAY
receive a request that has to be processed by your second Tomcat. So it
will connect to it, without dropping the connection to the first Tomcat.
So maxProcessors >= MaxClients is true for any scenario with Apache and
Tomcat.
Even worse...
If you have two Apache and two Tomcat,
maxProcessors >= sum(MaxClients) for all your Apache servers.
Otherwise, busy sites experience hangs.
Of course, if you do not use Apache as a frontend, and you simply have a
load balancer, you can keep maxProcessors much lower.
Yours,
Antonio Fiol
Pepijn Palmans - Kangaroot wrote:
>That is not true. This can be only true if you have only one instance of
>tomcat, and ALL (even static!!!) content is serverd by tomcat.
>
>Under normal setup you let your static content be served by apache, and
>only use tomcat when needed. So you might wanna have more apache
>processes.
>
>Pepijn
>
>On Fri, 20 Feb 2004 meissa.Sakho@assetmanagement.natexis.fr wrote:
>
>
>
>>As I know, the maxProcessor value must match
>>the MaxClients value in apache, because each apache
>>process must connect to a processor in Tomcat.
>>_______________________________________________
>>NATEXIS ASSET MANAGEMENT
>>Meissa SAKHO
>>01 58 19 45 71. . . . . . . . . . . . (84571)
>>meissa.sakho@assetmanagement.natexis.fr
>>
>>
Re: Réf. : Re: maxProcessors = max ammout ofconcurrent connections?
Posted by Pepijn Palmans - Kangaroot <pe...@kangaroot.net>.
That is not true. This can be only true if you have only one instance of
tomcat, and ALL (even static!!!) content is serverd by tomcat.
Under normal setup you let your static content be served by apache, and
only use tomcat when needed. So you might wanna have more apache
processes.
Pepijn
On Fri, 20 Feb 2004 meissa.Sakho@assetmanagement.natexis.fr wrote:
>
> As I know, the maxProcessor value must match
> the MaxClients value in apache, because each apache
> process must connect to a processor in Tomcat.
> _______________________________________________
> NATEXIS ASSET MANAGEMENT
> Meissa SAKHO
> 01 58 19 45 71. . . . . . . . . . . . (84571)
> meissa.sakho@assetmanagement.natexis.fr
>
>
>
>
> "Christian Hufgard" <ch...@gmx.de>
> 20/02/2004 10:03
> Veuillez r�pondre � "Tomcat Users List"
>
>
> Pour : "Tomcat Users List" <to...@jakarta.apache.org>
> cc :
> Objet : Re: maxProcessors = max ammout of concurrent connections?
>
>
> > I have several machines running tomcat now, but I like to get it tuned
> for
> > higher load. How can I find out the maximum processors I shoudl accept
> ?
> > Just putting in a high number of course results in very slow processing
> of
> > request, resulting in people refreshing, results in even more processes,
> > results in even slower replies, results in more refreshers :)
>
> Well, think you should not set the max processors to high, that the
> machine
> takes longer time to handle the requests. You should define your maximum
> response time and the figure out, how many request you can process at once
> in
> this time.
>
>
> > Any idea how to 'tune' the Maximum connectors and the acceptcount. Is
> it
> > really needed to have an acceptcount ?
> If the acceptCount is zero, the user will get an message, saying him that
> the server cannot be reached. If he is in the accepted queue, he will be
> shown
> that the server has been contacted and is about to process his request.
>
> Christian
>
>
>
>
> > Is there any document 'out in the wild' describing how to know which
> > values you should use ?
> >
> > I'm just getting to know tomcat but I find it hard to find the right
> > documentation. (many books describe the options in a complete
> > different/confising way).
> >
> > I have several machines running tomcat now, but I like to get it tuned
> for
> > higher load. How can I find out the maximum processors I shoudl accept
> ?
> > Just putting in a high number of course results in very slow processing
> of
> > request, resulting in people refreshing, results in even more processes,
> > results in even slower replies, results in more refreshers :)
> >
> > Any idea how to 'tune' the Maximum connectors and the acceptcount. Is
> it
> > really needed to have an acceptcount ?
> >
> > Pepijn
> >
> > On Fri, 20 Feb 2004, Christian Hufgard wrote:
> >
> > > > Why put minProcessors higher then the maximum allowed ?
> > >
> > > Think this does make no sense and should result in either a constant
> > ammount
> > > of processors (the min value, since it cannot fall lower), an error
> > message
> > > at startup or some other, unexpected behaviour.
> > > Now I set the minProcessors to 30, maxProcessor to 10.
> > > What happens is: No error messages in any log.
> > > 8 connections after startup of my prog.
> > > 9th connection after two seconds
> > > 10th connections after nine seconds.
> > > Then with each second a new connection is accepted. (Stopped the prog
> at
> > 20
> > > connections)
> > >
> > > After restarting 18 connections are established directly, two more
> after
> > > some seconds.
> > >
> > > 50 seconds after starting the program:
> > > Connections are interrupted. Logfile says:
> > > INFO: All threads are busy, waiting. Please increase maxThreads or
> check
> > the
> > > servlet status10 10.
> > >
> > >
> > >
> > >
> > > > I'm having trouble finding the exact meaning of some of the
> parameters
> > > > myself... but this might be part of a solution for your problem:
> > > >
> > > > Why put minProcessors higher then the maximum allowed ?
> > > >
> > > > Pepijn Palmans
> > > > Managing Director
> > > >
> > > > Kangaroot Linux Solutions
> > > > Grote Steenweg 91
> > > > 2600 Antwerp, Belgium
> > > > Tel: +32 3/286.17.17
> > > > Fax: +32 3/281.23.49
> > > >
> > > >
> > > > On Fri, 20 Feb 2004, Christian Hufgard wrote:
> > > >
> > > > > Hi folks,
> > > > >
> > > > > did not found this question answered, neither in the faqs, nor in
> > the
> > > > > archive of this mailing list, nor in bugzilla.
> > > > >
> > > > > Like in the subject described, my question is, wether the
> attribute
> > > > > maxProcessors in the Connector tag sets the maximum ammount of
> > > > concurrent connections
> > > > > the CoyeteConnector handles.
> > > > >
> > > > > I tested this on Tomcat up to 4.1.24 with jdk 1.4.2 on debian with
> > > > kernel
> > > > > 2.4.19, since this is our productive environment. If i set up the
> > > > connector in
> > > > > server.xml with the following parameters:
> > > > >
> > > > > <!-- Define a non-SSL Coyote HTTP/1.1 Connector on port 8080
> -->
> > > > > <Connector
> className="org.apache.coyote.tomcat4.CoyoteConnector"
> > > > > port="8080" minProcessors="20" maxProcessors="10"
> > > > > enableLookups="true" redirectPort="8443"
> > > > > acceptCount="5" debug="0" connectionTimeout="20000"
> > > > > useURIValidationHack="false"
> > disableUploadTimeout="true"
> > > > />
> > > > >
> > > > > I find a pretty strange behaviour (if maxProcessors should be max
> > > > > Connections):
> > > > > My test programm (jdk 1.4) tries to open up 80 connections to the
> > tomcat
> > > > and
> > > > > send some continueus requests to a webapps that create some load
> on
> > the
> > > > > server. The requests use HTTP/1.1 to keep the connection opened
> and
> > > > request a
> > > > > page each half second with variing parameters..
> > > > > I would expect, that 15 connections would be opened. 10 active + 5
> > on
> > > > hold.
> > > > > What I see is, that normally 18 connections are accepted upon
> > startup of
> > > > my
> > > > > programm. After 8 seconds another two connections are established,
> > 12
> > > > seconds
> > > > > later, again two new connections... This leads up to about 30-40
> > > > > connections.
> > > > >
> > > > > So, finally I precise my question a bit: What exactly does
> > maxProcessors
> > > > > mean? If it does not mean max concurrent connections, how many
> > > > connections can
> > > > > one processor handle?
> > > > >
> > > > > Hope I could describe my problem fairly enough :)
> > > > >
> > > > > Greets,
> > > > >
> > > > > Christian
> > > > >
> > > > > --
> > > > > GMX ProMail (250 MB Mailbox, 50 FreeSMS, Virenschutz, 2,99
> > EUR/Monat...)
> > > > > jetzt 3 Monate GRATIS + 3x DER SPIEGEL +++
> > http://www.gmx.net/derspiegel
> > > > +++
> > > > >
> > > > >
> > > > >
> > ---------------------------------------------------------------------
> > > > > To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
> > > > > For additional commands, e-mail:
> tomcat-user-help@jakarta.apache.org
> > > > >
> > > >
> > > >
> ---------------------------------------------------------------------
> > > > To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
> > > > For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
> > > >
> > >
> > > --
> > > GMX ProMail (250 MB Mailbox, 50 FreeSMS, Virenschutz, 2,99
> EUR/Monat...)
> > > jetzt 3 Monate GRATIS + 3x DER SPIEGEL +++ http://www.gmx.net/derspiegel
> > +++
> > >
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
> > > For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
> > >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
> > For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
> >
>
> --
> GMX ProMail (250 MB Mailbox, 50 FreeSMS, Virenschutz, 2,99 EUR/Monat...)
> jetzt 3 Monate GRATIS + 3x DER SPIEGEL +++ http://www.gmx.net/derspiegel +++
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
>
>
>
>
>
> L'integrite de ce message n'etant pas assuree sur internet, Natexis
> Banques Populaires ne peut etre tenu responsable de
> son contenu. Toute utilisation ou diffusion non autorisee est
> interdite. Si vous n'etes pas destinataire de ce message, merci de le
> detruire et d'avertir l'expediteur.
>
> The integrity of this message cannot be guaranteed
> on the Internet. Natexis Banques Populaires can not therefore be
> considered responsible for the contents.Any unauthorized use or dissemination is prohibited.
> If you are not the intended recipient of this message, then please delete it and
> notify the sender.
---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org