Re: mod_jk with jetty strange behaviour
Hello Rainer, I've let it log with debug log level all night and I managed to get what you needed (some [error] lines). I have the sample at http://www.craiova.ro/mod_jk.log Cheers, Andrei Neagoe Rainer Jung wrote: OK: you started with an error log entry referring to errno 104 as your original problem. I had the impression, you can easily reproduce this problem. In the logs you presented - even in the full log download - this problem is not present. The lines you referred to below are normal. They only tell us, that a new connection to the backend had to be established. Those log lines themselves do not indicate any problem. Regards, Rainer Andrei Neagoe wrote: Yes, they are. It was my mistake referring to it as an error. I was talking about: [Sun Aug 12 22:38:53 2007] [22781:3079018176] [debug] ajp_get_endpoint::jk_ajp_common.c (2498): acquired connection pool slot=0 [Sun Aug 12 22:38:53 2007] [22781:3079018176] [debug] ajp_marshal_into_msgb::jk_ajp_common.c (551): ajp marshaling done [Sun Aug 12 22:38:53 2007] [22781:3079018176] [debug] ajp_service::jk_ajp_common.c (1935): processing worker1 with 2 retries [Sun Aug 12 22:38:53 2007] [22781:3079018176] [debug] ajp_send_request::jk_ajp_common.c (1199): (worker1) socket 79 is not connected any more (errno=11) [Sun Aug 12 22:38:53 2007] [22781:3079018176] [debug] ajp_send_request::jk_ajp_common.c (1231): (worker1) failed sending request. Will try another pooled connection which I sent in the previous mail and you said it stopped right before something interesting. That's why I've attached the whole log. It doesn't seem anything suspicious to me around those log entries tough... Andrei Neagoe Rainer Jung wrote: Andrei Neagoe wrote: I've archived a portion of the log containing 2 instances of the error. You can download it at http://www.craiova.ro/mod_jk.log.bz2 I've tried sending it via mail but there's a limit on mail size. I downloaded this file, uncompressed and untarred. The resulting log file was -rw-r- 1 jung kippdata 9454912 Aug 12 21:39 mod_jk.log with md5 sum 4b880aa29a4a9baa478c543cd65dde3a. It does not contain any error! All log lines are info or debug. Regards, Rainer - To start a new topic, e-mail: users@tomcat.apache.org To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To start a new topic, e-mail: users@tomcat.apache.org To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: mod_jk with jetty strange behaviour
Yes, they are. It was my mistake referring to it as an error. I was talking about: [Sun Aug 12 22:38:53 2007] [22781:3079018176] [debug] ajp_get_endpoint::jk_ajp_common.c (2498): acquired connection pool slot=0 [Sun Aug 12 22:38:53 2007] [22781:3079018176] [debug] ajp_marshal_into_msgb::jk_ajp_common.c (551): ajp marshaling done [Sun Aug 12 22:38:53 2007] [22781:3079018176] [debug] ajp_service::jk_ajp_common.c (1935): processing worker1 with 2 retries [Sun Aug 12 22:38:53 2007] [22781:3079018176] [debug] ajp_send_request::jk_ajp_common.c (1199): (worker1) socket 79 is not connected any more (errno=11) [Sun Aug 12 22:38:53 2007] [22781:3079018176] [debug] ajp_send_request::jk_ajp_common.c (1231): (worker1) failed sending request. Will try another pooled connection which I sent in the previous mail and you said it stopped right before something interesting. That's why I've attached the whole log. It doesn't seem anything suspicious to me around those log entries tough... Andrei Neagoe Rainer Jung wrote: Andrei Neagoe wrote: I've archived a portion of the log containing 2 instances of the error. You can download it at http://www.craiova.ro/mod_jk.log.bz2 I've tried sending it via mail but there's a limit on mail size. I downloaded this file, uncompressed and untarred. The resulting log file was -rw-r- 1 jung kippdata 9454912 Aug 12 21:39 mod_jk.log with md5 sum 4b880aa29a4a9baa478c543cd65dde3a. It does not contain any error! All log lines are info or debug. Regards, Rainer - To start a new topic, e-mail: users@tomcat.apache.org To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To start a new topic, e-mail: users@tomcat.apache.org To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: mod_jk with jetty strange behaviour
I've archived a portion of the log containing 2 instances of the error. You can download it at http://www.craiova.ro/mod_jk.log.bz2 I've tried sending it via mail but there's a limit on mail size. Thanks, Andrei Neagoe Rainer Jung wrote: Your log excerpt stops exactly, when it's getting interesting. mod_jk connects and starts to send the request, no error until that point... We would like to check, at what point in the communication the problems appears, and what happened immediately before etc. Andrei Neagoe wrote: Thanks for your reply. I've tried looking around in jetty logs (also in webapps logs) but nothing related to the errors in mod_jk. Also I've tried to increase the log level... and I did not find anything (from my point of view): ... [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] ajp_connect_to_endpoint::jk_ajp_common.c (865): Connected socket 79 to (127.0.0.1:8019) [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] ajp_connection_tcp_send_message::jk_ajp_common.c (911): sending to ajp13 pos=4 len=514 max=8192 [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] ajp_connection_tcp_send_message::jk_ajp_common.c (911): 12 34 01 FE 02 02 00 08 48 54 54 50 2F 31 2E 31 - .4..HTTP/1.1 Thanks again, Andrei Neagoe - To start a new topic, e-mail: users@tomcat.apache.org To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To start a new topic, e-mail: users@tomcat.apache.org To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: mod_jk with jetty strange behaviour
Thanks for your reply. I've tried looking around in jetty logs (also in webapps logs) but nothing related to the errors in mod_jk. Also I've tried to increase the log level... and I did not find anything (from my point of view): [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] map_uri_to_worker::jk_uri_worker_map.c (597): Attempting to map URI '/craiova/images/utils/topnou.jpg' from 2 maps [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] map_uri_to_worker::jk_uri_worker_map.c (609): Attempting to map context URI '/craiova/*=worker1' source 'JkMount' [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] map_uri_to_worker::jk_uri_worker_map.c (624): Found a wildchar match '/craiova/*=worker1' [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] jk_handler::mod_jk.c (2131): Into handler jakarta-servlet worker=worker1 r->proxyreq=0 [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] wc_get_worker_for_name::jk_worker.c (115): found a worker worker1 [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] wc_get_name_for_type::jk_worker.c (292): Found worker type 'ajp13' [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] init_ws_service::mod_jk.c (811): Service protocol=HTTP/1.1 method=GET host=(null) addr=194.102.125.46 name=www.craiova.ro port=80 auth=(null) user=(null) laddr=217.156.64.186 raddr=194.102.125.46 uri=/craiova/images/utils/topnou.jpg [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] ajp_get_endpoint::jk_ajp_common.c (2498): acquired connection pool slot=0 [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] ajp_marshal_into_msgb::jk_ajp_common.c (551): ajp marshaling done [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] ajp_service::jk_ajp_common.c (1935): processing worker1 with 2 retries [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] ajp_send_request::jk_ajp_common.c (1199): (worker1) socket 79 is not connected any more (errno=11) [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] ajp_send_request::jk_ajp_common.c (1231): (worker1) failed sending request. Will try another pooled connection [Sun Aug 12 22:15:49 2007] [22077:3079018176] [info] ajp_send_request::jk_ajp_common.c (1265): (worker1) all endpoints are disconnected [Sun Aug 12 22:15:49 2007] [22077:3079018176] [info] ajp_service::jk_ajp_common.c (2085): (worker1) sending request to tomcat failed, recoverable operation attempt=1 [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] jk_open_socket::jk_connect.c (357): socket TCP_NODELAY set to On [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] jk_open_socket::jk_connect.c (370): socket SO_KEEPALIVE set to On [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] jk_open_socket::jk_connect.c (420): timeout 500 set for socket=79 [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] jk_open_socket::jk_connect.c (455): trying to connect socket 79 to 127.0.0.1:8019 [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] jk_open_socket::jk_connect.c (481): socket 79 connected to 127.0.0.1:8019 [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] ajp_connect_to_endpoint::jk_ajp_common.c (865): Connected socket 79 to (127.0.0.1:8019) [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] ajp_connection_tcp_send_message::jk_ajp_common.c (911): sending to ajp13 pos=4 len=514 max=8192 [Sun Aug 12 22:15:49 2007] [22077:3079018176] [debug] ajp_connection_tcp_send_message::jk_ajp_common.c (911): 12 34 01 FE 02 02 00 08 48 54 54 50 2F 31 2E 31 - .4..HTTP/1.1 Thanks again, Andrei Neagoe Rainer Jung wrote: No idea, what's happening, but if you can reproduce very easily, then you could increase the jk log level to debug or even trace. That way we will see more precisely, at what point the communication breaks. You error code 104 on many platforms is "Connection reset by peer". So there might be additional log information in some jetty log. Concerning the config (remarks meant for 1.2.25): >JkLogStampFormat "[%a %b %d %H:%M:%S %Y] " I think that's the default. If you simply delete it, you profit from the additional logging of milliseconds starting with 1.2.25. >JkOptions +ForwardKeySize +ForwardURICompat -ForwardDirectories Only use those, if you know you need them. Don't just copy tghem over from older configs, even if they used to be recommended. > workers.properties: > >worker.worker1.local_worker=1 Doesn't exist any more. Only useful when used in combination with a load balancer. Replaced by "distance", but since you don't use a JK load balancer simply delete it. >worker.worker1.lbfactor=1 Default value and only useful in combination with a JK load balancer. > worker.worker1.socket_timeout=500 500 seconds seems to be a relatively high timeout. No problem per se. Regards, Rainer Andrei Neagoe wrote: Hello, Until a rec
mod_jk with jetty strange behaviour
Hello, Until a recent upgrade I've been using Apache2 with Jetty and mod_jk2 just fine. Now, after I upgraded Apache to 2.2.3 and switched to mod_jk 1.2.18 and having the same version of Jetty (4.2.24) I find the following in the mod_jk.log: [Sun Aug 12 15:53:13 2007] [9685:3078895296] [error] ajp_connection_tcp_get_message::jk_ajp_common.c (966): (worker1) can't receive the response message from tomcat, network problems or tomcat (127.0.0.1:8019) is down (errno=104) [Sun Aug 12 15:53:13 2007] [9685:3078895296] [error] ajp_get_reply::jk_ajp_common.c (1658): (worker1) Tomcat is down or refused connection. No response has been sent to the client (yet) [Sun Aug 12 15:53:13 2007] [9685:3078895296] [info] ajp_service::jk_ajp_common.c (2046): (worker1) receiving from tomcat failed, recoverable operation attempt=0 [Sun Aug 12 15:53:13 2007] [9685:3078895296] [info] ajp_service::jk_ajp_common.c (2085): (worker1) sending request to tomcat failed, recoverable operation attempt=1 [Sun Aug 12 15:53:13 2007] [9685:3078895296] [error] ajp_connection_tcp_get_message::jk_ajp_common.c (966): (worker1) can't receive the response message from tomcat, network problems or tomcat (127.0.0.1:8019) is down (errno=104) [Sun Aug 12 15:53:13 2007] [9685:3078895296] [error] ajp_get_reply::jk_ajp_common.c (1658): (worker1) Tomcat is down or refused connection. No response has been sent to the client (yet) [Sun Aug 12 15:53:13 2007] [9685:3078895296] [info] ajp_service::jk_ajp_common.c (2046): (worker1) receiving from tomcat failed, recoverable operation attempt=1 [Sun Aug 12 15:53:13 2007] [9685:3078895296] [info] ajp_service::jk_ajp_common.c (2085): (worker1) sending request to tomcat failed, recoverable operation attempt=2 [Sun Aug 12 15:53:13 2007] [9685:3078895296] [error] ajp_service::jk_ajp_common.c (2097): (worker1) Connecting to tomcat failed. Tomcat is probably not started or is listening on the wrong port [Sun Aug 12 15:53:13 2007] [9685:3078895296] [info] jk_handler::mod_jk.c (2277): Service error=0 for worker=worker1 Both Apache and Jetty are running on the same box. mod_jk.conf looks like this: JkWorkersFile /etc/apache2/workers.properties JkLogFile /var/log/apache2/mod_jk.log JkLogLevelinfo JkLogStampFormat "[%a %b %d %H:%M:%S %Y] " JkOptions +ForwardKeySize +ForwardURICompat -ForwardDirectories JkShmFile /var/log/apache2/jk.shm workers.properties: worker.list=worker1 worker.worker1.port=8019 worker.worker1.host=127.0.0.1 worker.worker1.type=ajp13 worker.worker1.local_worker=1 worker.worker1.retries=2 worker.worker1.lbfactor=1 worker.worker1.socket_timeout=500 worker.worker1.socket_keepalive=1 Server behaviour: everything seems to be working fine except the admin module of a certain servlet, which gives an http error (don't know exactly as I don't have access to it) but after 1-2 refreshes workes fine. I've also tried to switch to latest mod_jk, 1.2.25 (above logs are from 1.2.25) but same thing. What I can be sure of is that jetty's running just fine, regular requests are served without ANY problems. I've browsed through a servlet for at least half an hour and no error. Client feedback also confirms no weird behavior. After googleing for a couple of days and trying different workers.properties settings, I still have no clue on what might be the cause of those log entries. Any suggestion is appreciated. Thanks, -- Andrei Neagoe