Hi, Thanks for the reply.
Here are my answers. OS: RHEL 5 S/w: Apache 2.2.15, Mod_jk 1.2.x, Jboss 4.2.1GA - Yes, I am able to reproduce the situation ( actually this situation is consistent not going away) - I cannot check the traffic from the same machine as the machine located in a datacenter. - I even modified the firewall rules to block access to Port 80 to the whole world and allow only single IP of my network. Still the problem persists. - No Errors were reproted at Apache level - When Apache is on a different machine the functionality looks normal. - I do suspect the reason give by you is valid that some network device is playing a role in between to close the traffic. Let me confirm that also. Please send any more suggestions. Regards, Srinu. On Thu, Jul 15, 2010 at 8:59 PM, Rainer Jung <rainer.j...@kippdata.de>wrote: > On 15.07.2010 15:23, Rocky Chak wrote: > >> Hi, >> >> I have a application using Apache 2.2.15, Mod_jk and Jboss 4.2.1. Apache >> and Jboss Running fine till now. Suddenly there is an issue of slow >> browsing and later it stopped serving pages. >> >> We moved to backup server and now trying to fix this one. >> >> When Application deployed on Jboss and Jboss run on Port 80 .. Application >> access& functionality is PERFECT. >> >> When Apache is on a different machine with same configuration of the >> present >> host with mod_jk ... .. Application access& functionality is PERFECT. >> >> But when Apache and Jboss are configured on the same machine with Jboss on >> 8080 ( Anyway this port is not required as AJP uses 8009 port ) - >> Application is behaving in weird fashion. JSP Pages in the application >> are >> not rendered fully. No error is displayed in Apache access_log. >> >> But mod_jk.log shows these errors: >> >> [Thu Jul 15 08:24:39 2010][29201:3086371680] [info] >> ajp_process_callback::jk_ajp_common.c (1788): Writing to client aborted or >> client network problems >> [Thu Jul 15 08:24:39 2010][29201:3086371680] [info] >> ajp_service::jk_ajp_common.c (2447): (node1) sending request to tomcat >> failed (unrecoverable), because of client write error (attempt=1) >> [Thu Jul 15 08:24:41 2010][29201:3086371680] [info] >> service::jk_lb_worker.c >> (1384): service failed, worker node1 is in local error state >> [Thu Jul 15 08:24:41 2010][29201:3086371680] [info] >> service::jk_lb_worker.c >> (1403): unrecoverable error 200, request failed. Client failed in the >> middle >> of request, we can't recover to another instance. >> [Thu Jul 15 08:24:41 2010]loadbalancer abc.xyz.com 70.790837 (url removed >> by >> me) >> [Thu Jul 15 08:24:41 2010][29201:3086371680] [info] jk_handler::mod_jk.c >> (2608): Aborting connection for worker=loadbalancer >> [Thu Jul 15 08:24:48 2010][29199:3086371680] [info] >> ajp_process_callback::jk_ajp_common.c (1788): Writing to client aborted or >> client network problems >> [Thu Jul 15 08:24:48 2010][29199:3086371680] [info] >> ajp_service::jk_ajp_common.c (2447): (node1) sending request to tomcat >> failed (unrecoverable), because of client write error (attempt=1) >> [Thu Jul 15 08:24:50 2010][29199:3086371680] [info] >> service::jk_lb_worker.c >> (1384): service failed, worker node1 is in local error state >> [Thu Jul 15 08:24:50 2010][29199:3086371680] [info] >> service::jk_lb_worker.c >> (1403): unrecoverable error 200, request failed. Client failed in the >> middle >> of request, we can't recover to another instance. >> [Thu Jul 15 08:24:50 2010]loadbalancer abc.xyz.com 10.869269 >> [Thu Jul 15 08:24:50 2010][29199:3086371680] [info] jk_handler::mod_jk.c >> (2608): Aborting connection for worker=loadbalancer >> [Thu Jul 15 08:24:52 2010][29195:3086371680] [info] >> ajp_process_callback::jk_ajp_common.c (1788): Writing to client aborted or >> client network problems >> [Thu Jul 15 08:24:52 2010][29195:3086371680] [info] >> ajp_service::jk_ajp_common.c (2447): (node1) sending request to tomcat >> failed (unrecoverable), because of client write error (attempt=1) >> [Thu Jul 15 08:24:54 2010][29195:3086371680] [info] >> service::jk_lb_worker.c >> (1384): service failed, worker node1 is in local error state >> [Thu Jul 15 08:24:54 2010][29195:3086371680] [info] >> service::jk_lb_worker.c >> (1403): unrecoverable error 200, request failed. Client failed in the >> middle >> of request, we can't recover to another instance. >> [Thu Jul 15 08:24:54 2010]loadbalancer abc.xyz.com 6.253777 >> [Thu Jul 15 08:24:54 2010][29195:3086371680] [info] jk_handler::mod_jk.c >> (2608): Aborting connection for worker=loadbalancer >> [Thu Jul 15 08:24:55 2010][29198:3086371680] [info] >> ajp_process_callback::jk_ajp_common.c (1788): Writing to client aborted or >> client network problems >> [Thu Jul 15 08:24:55 2010][29198:3086371680] [info] >> ajp_service::jk_ajp_common.c (2447): (node1) sending request to tomcat >> failed (unrecoverable), because of client write error (attempt=1) >> [Thu Jul 15 08:24:57 2010][29198:3086371680] [info] >> service::jk_lb_worker.c >> (1384): service failed, worker node1 is in local error state >> [Thu Jul 15 08:24:57 2010][29198:3086371680] [info] >> service::jk_lb_worker.c >> (1403): unrecoverable error 200, request failed. Client failed in the >> middle >> of request, we can't recover to another instance. >> >> This error occurs whenever there is a request. >> >> Can somebody please help why this is happening. >> > > What's your operating system? > > The error means, that mod_jk wants to send back response packet via Apache > and Apache told mod_jk that it couldn't send the data to the client/browser. > Usually that means the connections has been closed by the browser. > > Occasional occurence of this (e.g. for 1 percent of all requests) is OK, > because typically it happens, when a user doesn't wait for the full answer > and instead proceeds clicking. Then the browser closes the previous > connection, likely with a connection reset, and starts a new connection for > the next request. The web server doesn't know about the closed connection > and only when it tries to send the next packet it gets an error from the OS, > that the connction is already closed. > > 1) Look at your Apache error log for related errors. > > 2) Can you reproduce the problem? If so: > - does it also happen, if you run the client on the same machine as the > server? Any other depenency on the client network? > - sniff all network traffic between the Apache port and your client used > for reproduction (not using the local host in this case). Look whether the > client actually closed the connection prematurely. If you are using HTTP and > not HTTPS, then you can easily see from the response packets, whether the > response was finished or not. > > Note: we recently had a customr who observed the same problem. It turned > out, that indeed reset packets were send to the web erver. Checking the MAC > addresses revealed, those packets were not coming rom the browser, but > instead from some other network security device. Thi device decided the > traffic was malicious and send a reset packet. The traffic was not going > through the device, it simply sniffed the LAN traffic and inserted a reset > packet when it thought it would help security ... > > Regards, > > Rainer > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org > For additional commands, e-mail: users-h...@tomcat.apache.org > >