You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@tomcat.apache.org by heind heind <lo...@gmail.com> on 2007/01/23 17:49:29 UTC

mod_jk : error sending request. Will try another pooled connection

Hi,

I have a web application deployed in tomcat embedded in jonas 4.7.4
Application Server, and  in frontend  there are apache 2.0.55 and mod_jk
12.20.

My issue is that the message below is generated systematically after every
jsp call although the jsp are well served and the application works fine.


[2007-01-18 14:51:48 (Thu 18 Jan 2007)] [19047:0001] [info]
ajp_send_request::jk_ajp_common.c (1201): (mgf_worker2) error sending
request. Will try another p
ooled connection
[2007-01-18 14:51:48 (Thu 18 Jan 2007)] [19047:0001] [info]
ajp_send_request::jk_ajp_common.c (1227): (mgf_worker2) all endpoints are
disconnected
[2007-01-18 14:51:48 (Thu 18 Jan 2007)] [19047:0001] [info]
ajp_send_request::jk_ajp_common.c (1230): (mgf_worker2) increase the backend
idle connection time
out or the connection_pool_minsize
[2007-01-18 14:51:48 (Thu 18 Jan 2007)] [19047:0001] [info]
ajp_service::jk_ajp_common.c (1916): (mgf_worker2) sending request to tomcat
failed,  recoverable
 operation attempt=1
[2007-01-18 14:53:09 (Thu 18 Jan 2007)] [19053:0001] [info]
ajp_send_request::jk_ajp_common.c (1201): (mgf_worker2) error sending
request. Will try another p
ooled connection
[2007-01-18 14:53:09 (Thu 18 Jan 2007)] [19053:0001] [info]
ajp_send_request::jk_ajp_common.c (1227): (mgf_worker2) all endpoints are
disconnected
[2007-01-18 14:53:09 (Thu 18 Jan 2007)] [19053:0001] [info]
ajp_send_request::jk_ajp_common.c (1230): (mgf_worker2) increase the backend
idle connection time
out or the connection_pool_minsize
[2007-01-18 14:53:09 (Thu 18 Jan 2007)] [19053:0001] [info]
ajp_service::jk_ajp_common.c (1916): (mgf_worker2) sending request to tomcat
failed,  recoverable
 operation attempt=1

My question is what is the meaning of this message and how to avoid a second
attempt ?

PS :
- my Apache is 2.0.55 prefork so I can't increase the connectio_pool_minsize
(cf. http://tomcat.apache.org/connectors-doc/reference/workers.html )
- the workers.properties is in standard configuration.


Thanks,

Hind

Re: mod_jk : error sending request. Will try another pooled connection

Posted by heind heind <lo...@gmail.com>.
Hi.

Thanks for answering me.


On 1/23/07, Rainer Jung <ra...@kippdata.de> wrote:
>
> Hi,
>
> heind heind schrieb:
> > Hi,
> >
> > I have a web application deployed in tomcat embedded in jonas 4.7.4
> > Application Server, and  in frontend  there are apache 2.0.55 and mod_jk
> > 12.20.
> >
> > My issue is that the message below is generated systematically after
> every
> > jsp call although the jsp are well served and the application works
> fine.
> >
> >
>
> ...
>
> > My question is what is the meaning of this message and how to avoid a
> > second
> > attempt ?
> >
> > PS :
> > - my Apache is 2.0.55 prefork so I can't increase the
> > connectio_pool_minsize
> > (cf. http://tomcat.apache.org/connectors-doc/reference/workers.html )
> > - the workers.properties is in standard configuration.
>
> >There is no "standard" configuration. Please post your
> >workers.properties, the Jk* directives and the config of the prefork MPM
> >from your apache config and the tomcat AJP/1.3 connector config from
> >server.xml.


workers.properties
worker.list=w,jkstatus
# workers vers v440r-01
worker.w_worker1.port=8407
worker.w_worker1.host=v440r-01
worker.w_worker1.type=ajp13
worker.w_worker1.lbfactor=1
worker.w_worker1.distance=1
worker.w_worker1.connection_pool_minsize=1
worker.w_worker1.connection_pool_size=2

# workers vers v440r-02

worker.w_worker2.port=8407
worker.w_worker2.host=10.198.45.5
worker.w_worker2.type=ajp13
worker.w_worker2.lbfactor=1
worker.w_worker2.distance=0
worker.w_worker2.connection_pool_minsize=1
worker.w_worker2.connection_pool_size=2
# LB worker
worker.w.type=lb
worker.w.balance_workers=w_worker2
#worker.w.balance_workers=w_worker1,w_worker2
worker.jkstatus.type=status


jk directives

JkWorkersFile /jonasintra/logiciel/apache_mgf/conf/workers.properties
JkLogFile "|/jonasintra/logiciel/apache_mgf/bin/rotatelogs -l
/jonasintra/logiciel/apache_mgf/logs/mod_jk_%Y-%m-%d-%Hh%M.log 86400"
JkLogLevel INFO
#JkLogStampFormat "[%a %b %d %H:%M:%S %Y]"
JkLogStampFormat "[%Y-%m-%d %T (%a %d %b %Y)] "
JkOptions +ForwardKeySize +ForwardURICompat -ForwardDirectories
#JkRequestLogFormat "%w %V %T"
JkRequestLogFormat "URL : %U - Query : %q - Duration : %Ts"
JkMount /mgf/clamweb.applications.gip.f1.ImageServlet mgf
JkMount /mgf/RefreshData mgf
JkMount /mgf/*.do mgf
JkMount /mgf/*.jsp mgf
JkMount /mgf_mrtg/*.jsp mgf
JkMount /jkmanager/* jkstatus
JkMount /mgfJonasAdmin/* mgf
JkMount /essais/* mgf

prefork mpm

<IfModule prefork.c>
ServerLimit         500
StartServers        20
MinSpareServers     10
MaxSpareServers     20
MaxClients          500
MaxRequestsPerChild 10000
</IfModule>

AJP/1.3 connector
  <Connector port="8407"
       maxThreads="500"
         enableLookups="false" redirectPort="9043" protocol="AJP/1.3"
connectionTimeout="600000"/>

 I set the connectionTimeout to the value above to avoide the broken pipe
messages in the backend.

>As long as there are no messages of higher level, those info messages do
> >not indicate a serious functional problem. Normally they should not show
> >up for each request. Are you sure that you get those for every request?


Yes, I am sure that this message shows up for evry request  when the time
between each call is about 5 min as you see below within the exemple from
mod_jk logs.

[2007-01-24 01:53:07 (Wed 24 Jan 2007)] [12965:0001] [info]
ajp_send_request::jk_ajp_common.c (1201): (mgf_worker2) error sending
request. Will try another pooled connection
[2007-01-24 01:53:07 (Wed 24 Jan 2007)] [12965:0001] [info]
ajp_send_request::jk_ajp_common.c (1227): (mgf_worker2) all endpoints are
disconnected
[2007-01-24 01:53:07 (Wed 24 Jan 2007)] [12965:0001] [info]
ajp_send_request::jk_ajp_common.c (1230): (mgf_worker2) increase the backend
idle connection timeout or the connection_pool_minsize
[2007-01-24 01:53:07 (Wed 24 Jan 2007)] [12965:0001] [info]
ajp_service::jk_ajp_common.c (1916): (mgf_worker2) sending request to tomcat
failed,  recoverable operation attempt=1
[2007-01-24 01:53:07 (Wed 24 Jan 2007)] URL : /mgf_mrtg/memory.jsp - Query
:  - Duration : 0.006798s

[2007-01-24 01:58:08 (Wed 24 Jan 2007)] [12841:0001] [info]
ajp_send_request::jk_ajp_common.c (1201): (mgf_worker2) error sending
request. Will try another pooled connection
[2007-01-24 01:58:08 (Wed 24 Jan 2007)] [12841:0001] [info]
ajp_send_request::jk_ajp_common.c (1227): (mgf_worker2) all endpoints are
disconnected
[2007-01-24 01:58:08 (Wed 24 Jan 2007)] [12841:0001] [info]
ajp_send_request::jk_ajp_common.c (1230): (mgf_worker2) increase the backend
idle connection timeout or the connection_pool_minsize
[2007-01-24 01:58:08 (Wed 24 Jan 2007)] [12841:0001] [info]
ajp_service::jk_ajp_common.c (1916): (mgf_worker2) sending request to tomcat
failed,  recoverable operation attempt=1
[2007-01-24 01:58:08 (Wed 24 Jan 2007)] URL : /mgf_mrtg/memory.jsp - Query
:  - Duration : 0.006775s

[2007-01-24 02:03:07 (Wed 24 Jan 2007)] [12792:0001] [info]
ajp_send_request::jk_ajp_common.c (1201): (mgf_worker2) error sending
request. Will try another pooled connection
[2007-01-24 02:03:07 (Wed 24 Jan 2007)] [12792:0001] [info]
ajp_send_request::jk_ajp_common.c (1227): (mgf_worker2) all endpoints are
disconnected
[2007-01-24 02:03:07 (Wed 24 Jan 2007)] [12792:0001] [info]
ajp_send_request::jk_ajp_common.c (1230): (mgf_worker2) increase the backend
idle connection timeout or the connection_pool_minsize
[2007-01-24 02:03:07 (Wed 24 Jan 2007)] [12792:0001] [info]
ajp_service::jk_ajp_common.c (1916): (mgf_worker2) sending request to tomcat
failed,  recoverable operation attempt=1
[2007-01-24 02:03:07 (Wed 24 Jan 2007)] URL : /mgf_mrtg/memory.jsp - Query
:  - Duration : 0.007115s

[2007-01-24 02:08:07 (Wed 24 Jan 2007)] [12952:0001] [info]
ajp_send_request::jk_ajp_common.c (1201): (mgf_worker2) error sending
request. Will try another pooled connection
[2007-01-24 02:08:07 (Wed 24 Jan 2007)] [12952:0001] [info]
ajp_send_request::jk_ajp_common.c (1227): (mgf_worker2) all endpoints are
disconnected
[2007-01-24 02:08:07 (Wed 24 Jan 2007)] [12952:0001] [info]
ajp_send_request::jk_ajp_common.c (1230): (mgf_worker2) increase the backend
idle connection timeout or the connection_pool_minsize
[2007-01-24 02:08:07 (Wed 24 Jan 2007)] [12952:0001] [info]
ajp_service::jk_ajp_common.c (1916): (mgf_worker2) sending request to tomcat
failed,  recoverable operation attempt=1
[2007-01-24 02:08:07 (Wed 24 Jan 2007)] URL : /mgf_mrtg/memory.jsp - Query
:  - Duration : 0.006293s
[2007-01-24 02:13:07 (Wed 24 Jan 2007




>What happens, if you use the binary "ab" from your apache httpd to send
> >let's say 1000 requests with a parallelity of 10 (ab -c 10 -n 1000
> >http://myserver:myport/myurl)?



Executing  ab -c 10 -n 1000  http://myserver:myport/essais/essais.jsp

Fisrt request
[2007-01-24 15:17:50 (Wed 24 Jan 2007)] URL : /essais/essais.jsp - Query :
- Duration : 10.298479s

Fisrt message in mod_jk logs (~10 min after)
[2007-01-24 15:27:36 (Wed 24 Jan 2007)] [12811:0001] [info]
ajp_send_request::jk_ajp_common.c (1227): (mgf_worker2) all endpoints are
disconnected
[2007-01-24 15:27:36 (Wed 24 Jan 2007)] [12811:0001] [info]
ajp_send_request::jk_ajp_common.c (1230): (mgf_worker2) increase the backend
idle connection time
out or the connection_pool_minsize
[2007-01-24 15:27:36 (Wed 24 Jan 2007)] [12811:0001] [info]
ajp_service::jk_ajp_common.c (1916): (mgf_worker2) sending request to tomcat
failed,  recoverable
 operation attempt=1

Second message (~5 min after the previous message)
[2007-01-24 15:32:44 (Wed 24 Jan 2007)] [12882:0001] [info]
ajp_process_callback::jk_ajp_common.c (1433): Writing to client aborted or
client network problem
s
[2007-01-24 15:32:44 (Wed 24 Jan 2007)] [12882:0001] [info]
ajp_service::jk_ajp_common.c (1832): (mgf_worker2) request failed, because
of client write error
without recovery in send loop attempt=0
[2007-01-24 15:32:44 (Wed 24 Jan 2007)] [12882:0001] [info]
service::jk_lb_worker.c (979): unrecoverable error 200, request failed.
Client failed in the midd
le of request, we can't recover to another instance.
[2007-01-24 15:32:44 (Wed 24 Jan 2007)] URL : /essais/essais.jsp - Query :
- Duration : 4.140661s
[2007-01-24 15:32:44 (Wed 24 Jan 2007)] [6187:0001] [info]
ajp_process_callback::jk_ajp_common.c (1433): Writing to client aborted or
client network problems
[2007-01-24 15:32:44 (Wed 24 Jan 2007)] [12882:0001] [info]
jk_handler::mod_jk.c (2135): Aborting connection for worker=mgf


The third message is different

[2007-01-24 15:32:44 (Wed 24 Jan 2007)] [12945:0001] [info]
service::jk_lb_worker.c (979): unrecoverable error 200, request failed.
Client failed in the middle of request, we can't recover to another
instance.
[2007-01-24 15:32:44 (Wed 24 Jan 2007)] URL : /essais/essais.jsp - Query :
- Duration : 16.967256s
[2007-01-24 15:32:44 (Wed 24 Jan 2007)] URL : /essais/essais.jsp - Query :
- Duration : 4.492996s


Regards,

PS : my apache is 2.0.59

Hind

Re: mod_jk : error sending request. Will try another pooled connection

Posted by Rainer Jung <ra...@kippdata.de>.
Hi,

heind heind schrieb:
> Hi,
> 
> I have a web application deployed in tomcat embedded in jonas 4.7.4
> Application Server, and  in frontend  there are apache 2.0.55 and mod_jk
> 12.20.
> 
> My issue is that the message below is generated systematically after every
> jsp call although the jsp are well served and the application works fine.
> 
> 

...

 > My question is what is the meaning of this message and how to avoid a
> second
> attempt ?
> 
> PS :
> - my Apache is 2.0.55 prefork so I can't increase the
> connectio_pool_minsize
> (cf. http://tomcat.apache.org/connectors-doc/reference/workers.html )
> - the workers.properties is in standard configuration.

There is no "standard" configuration. Please post your
workers.properties, the Jk* directives and the config of the prefork MPM
from your apache config and the tomcat AJP/1.3 connector config from
server.xml.

As long as there are no messages of higher level, those info messages do
not indicate a serious functional problem. Normally they should not show
up for each request. Are you sure that you get those for every request?
What happens, if you use the binary "ab" from your apache httpd to send
let's say 1000 requests with a parallelity of 10 (ab -c 10 -n 1000
http://myserver:myport/myurl)?

Regards,

Rainer

---------------------------------------------------------------------
To start a new topic, e-mail: users@tomcat.apache.org
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org