jk经常死掉，看了一下mod_jk.log，大家帮忙看看。

liwei6998

公司网站这几天经常不能访问。ssh进去后，看了看mod_jk.log，发现这些错误：

[Tue Oct 24 18:42:01 2006] worker1 219.234.80.176 0.000353
[Tue Oct 24 18:42:01 2006] [5595:52064] [info]  jk_handler::mod_jk.c (197: Service error=0 for worker=worker1
[Tue Oct 24 18:42:07 2006] [5621:52064] [info]  jk_open_socket::jk_connect.c (445): connect to 127.0.0.1:8009 failed with errno=111
[Tue Oct 24 18:42:07 2006] [5621:52064] [info]  ajp_connect_to_endpoint::jk_ajp_common.c (889): Failed opening socket to (127.0.0.1:8009) with
(errno=111)
[Tue Oct 24 18:42:07 2006] [5621:52064] [info]  ajp_send_request::jk_ajp_common.c (1251): Error connecting to the backend server.
[Tue Oct 24 18:42:07 2006] [5621:52064] [info]  ajp_service::jk_ajp_common.c (1783): Sending request to tomcat failed,  recoverable operation
attempt=1
[Tue Oct 24 18:42:07 2006] [5621:52064] [info]  jk_open_socket::jk_connect.c (445): connect to 127.0.0.1:8009 failed with errno=111
[Tue Oct 24 18:42:07 2006] [5621:52064] [info]  ajp_connect_to_endpoint::jk_ajp_common.c (889): Failed opening socket to (127.0.0.1:8009) with
(errno=111)
[Tue Oct 24 18:42:07 2006] [5621:52064] [info]  ajp_send_request::jk_ajp_common.c (1251): Error connecting to the backend server.
[Tue Oct 24 18:42:07 2006] [5621:52064] [info]  ajp_service::jk_ajp_common.c (1783): Sending request to tomcat failed,  recoverable operation
attempt=2
[Tue Oct 24 18:42:07 2006] [5621:52064] [error] ajp_service::jk_ajp_common.c (1794): Error connecting to tomcat. Tomcat is probably not starte
d or is listening on the wrong port. worker=worker1 failed
[Tue Oct 24 18:42:07 2006] worker1 mall.haohao.com 0.000414
[Tue Oct 24 18:42:07 2006] [5621:52064] [info]  jk_handler::mod_jk.c (197: Service error=0 for worker=worker1
[Tue Oct 24 18:42:08 2006] [5625:52064] [info]  ajp_send_request::jk_ajp_common.c (117: Socket 29 is not connected any more (errno=-1)
[Tue Oct 24 18:42:08 2006] [5625:52064] [info]  ajp_send_request::jk_ajp_common.c (1202): Error sending request. Will try another pooled conne
ction
[Tue Oct 24 18:42:08 2006] [5625:52064] [info]  ajp_send_request::jk_ajp_common.c (1224): All endpoints are disconnected or dead
[Tue Oct 24 18:42:08 2006] [5625:52064] [info]  ajp_service::jk_ajp_common.c (1783): Sending request to tomcat failed,  recoverable operation
attempt=1
[Tue Oct 24 18:42:08 2006] [5625:52064] [info]  jk_open_socket::jk_connect.c (445): connect to 127.0.0.1:8009 failed with errno=111
[Tue Oct 24 18:42:08 2006] [5625:52064] [info]  ajp_connect_to_endpoint::jk_ajp_common.c (889): Failed opening socket to (127.0.0.1:8009) with
(errno=111)
[Tue Oct 24 18:42:08 2006] [5625:52064] [info]  ajp_send_request::jk_ajp_common.c (1251): Error connecting to the backend server.
[Tue Oct 24 18:42:08 2006] [5625:52064] [info]  ajp_service::jk_ajp_common.c (1783): Sending request to tomcat failed,  recoverable operation
attempt=2
[Tue Oct 24 18:42:08 2006] [5625:52064] [error] ajp_service::jk_ajp_common.c (1794): Error connecting to tomcat. Tomcat is probably not starte
d or is listening on the wrong port. worker=worker1 failed
[Tue Oct 24 18:42:08 2006] worker1 www.haohao.com 0.000321
[Tue Oct 24 18:42:08 2006] [5625:52064] [info]  jk_handler::mod_jk.c (197: Service error=0 for worker=worker1
[Tue Oct 24 18:42:10 2006] [5583:52064] [info]  jk_open_socket::jk_connect.c (445): connect to 127.0.0.1:8009 failed with errno=111
[Tue Oct 24 18:42:10 2006] [5583:52064] [info]  ajp_connect_to_endpoint::jk_ajp_common.c (889): Failed opening socket to (127.0.0.1:8009) with
(errno=111)
[Tue Oct 24 18:42:10 2006] [5583:52064] [info]  ajp_send_request::jk_ajp_common.c (1251): Error connecting to the backend server.
[Tue Oct 24 18:42:10 2006] [5583:52064] [info]  ajp_service::jk_ajp_common.c (1783): Sending request to tomcat failed,  recoverable operation
attempt=1
[Tue Oct 24 18:42:10 2006] [5583:52064] [info]  jk_open_socket::jk_connect.c (445): connect to 127.0.0.1:8009 failed with errno=111
[Tue Oct 24 18:42:10 2006] [5583:52064] [info]  ajp_connect_to_endpoint::jk_ajp_common.c (889): Failed opening socket to (127.0.0.1:8009) with
(errno=111)
[Tue Oct 24 18:42:10 2006] [5583:52064] [info]  ajp_send_request::jk_ajp_common.c (1251): Error connecting to the backend server.
[Tue Oct 24 18:42:10 2006] [5583:52064] [info]  ajp_service::jk_ajp_common.c (1783): Sending request to tomcat failed,  recoverable operation
attempt=2
[Tue Oct 24 18:42:10 2006] [5583:52064] [error] ajp_service::jk_ajp_common.c (1794): Error connecting to tomcat. Tomcat is probably not starte
d or is listening on the wrong port. worker=worker1 failed


重启tomcat后，就可以了。

请大家帮忙看看，是什么原因？ 先谢了。

patirot

我也想知道原因，顶上去

KissNoway

是tomcat宕了，应该查tomcat的日至。

HonestQiao

楼上正确。

或者可以设置crontab，定期重启tomcat

everlasting_188

真是tomcat的问题,那就晕了,我也有这个问题

xpingtn

今天我也刚遇到这种错误，不知道怎么解决。。
不知道是什么原因导致tomcat错误。。。。。。

ipaddr

首先检查Tomcat是否正常，

我以前测试过，mod_jk确定不稳定，并发连接多的话，就会有些连接是断掉的。

后面是用mod_proxy

linuxpf

发现同样的问题,如果先启动apache,后tomcat,mod_jk.log报错:

[Sat Dec 20 16:38:10.562 2008] [2887:3086804672] [info] jk_handler::mod_jk.c (2476): Service error=-3 for worker=ajp1
[Mon Dec 22 07:50:42.259 2008] [2883:3086165696] [info] init_jk::mod_jk.c (3018): mod_jk/1.2.27 initialized
[Mon Dec 22 07:50:42.586 2008] [2889:3086165696] [info] init_jk::mod_jk.c (3018): mod_jk/1.2.27 initialized
[Mon Dec 22 08:41:52.555 2008] [2905:3086165696] [info] ajp_send_request::jk_ajp_common.c (1456): (ajp1) all endpoints are disconnected, detec
ted by connect check (1), cping (0), send (0)
[Mon Dec 22 08:41:52.868 2008] [2904:3086165696] [info] ajp_send_request::jk_ajp_common.c (1456): (ajp1) all endpoints are disconnected, detec
ted by connect check (1), cping (0), send (0)
[Mon Dec 22 08:41:52.953 2008] [2894:3086165696] [info] ajp_send_request::jk_ajp_common.c (1456): (ajp1) all endpoints are disconnected, detec
ted by connect check (1), cping (0), send (0)
[Mon Dec 22 08:42:06.667 2008] [2897:3086165696] [info] ajp_send_request::jk_ajp_common.c (1456): (ajp1) all endpoints are disconnected, detec
ted by connect check (1), cping (0), send (0)
[Mon Dec 22 08:42:07.030 2008] [2903:3086165696] [info] ajp_send_request::jk_ajp_common.c (1456): (ajp1) all endpoints are disconnected, detec
ted by connect check (1), cping (0), send (0)
[Mon Dec 22 08:55:59.869 2008] [2889:3086165696] [info] init_jk::mod_jk.c (3018): mod_jk/1.2.27 initialized
[Mon Dec 22 12:50:04.412 2008] [5313:3086939840] [info] init_jk::mod_jk.c (3018): mod_jk/1.2.27 initialized
[Mon Dec 22 12:50:04.669 2008] [5314:3086939840] [info] init_jk::mod_jk.c (3018): mod_jk/1.2.27 initialized
[Mon Dec 22 14:58:55.548 2008] [5318:3086939840] [info] ajp_process_callback::jk_ajp_common.c (1748): Writing to client aborted or client netw
ork problems
[Mon Dec 22 14:58:55.550 2008] [5318:3086939840] [info] ajp_service::jk_ajp_common.c (2407): (ajp1) sending request to tomcat failed (unrecove
rable), because of client write error (attempt=1)
[Mon Dec 22 14:58:55.558 2008] [5318:3086939840] [info] jk_handler::mod_jk.c (2469): Aborting connection for worker=ajp1
[Mon Dec 22 15:01:15.580 2008] [5318:3086939840] [info] ajp_process_callback::jk_ajp_common.c (1748): Writing to client aborted or client netw
ork problems
[Mon Dec 22 15:01:15.580 2008] [5318:3086939840] [info] ajp_service::jk_ajp_common.c (2407): (ajp1) sending request to tomcat failed (unrecove
rable), because of client write error (attempt=1)
[Mon Dec 22 15:01:15.584 2008] [5318:3086939840] [info] jk_handler::mod_jk.c (2469): Aborting connection for worker=ajp1
[Mon Dec 22 15:19:57.536 2008] [3683:3086616256] [info] init_jk::mod_jk.c (3018): mod_jk/1.2.27 initialized
[Mon Dec 22 15:19:57.803 2008] [3684:3086616256] [info] init_jk::mod_jk.c (3018): mod_jk/1.2.27 initialized

等一会就没有问题
使用apache 2.0.52
mod_jk 1.2.27

zawecha1

可以在workers.properties里设置相应的负载机器的recovery_options=3就可以了。
注解：
recovery_options属性说明了web server在检测到Tomcat失败后如何进行恢复工作。默认情况下，web server将转发请求给处于负载平衡模式中的另一个Tomcat。属性值为0，说明全部恢复；属性值为1，说明如果在Tomcat接到请求后出现失败状况，则不进行恢复；属性值为2，说明如果在Tomcat发送http头给客户端后出现失败状况，则不进行恢复；属性值为3，说明如果在Tomcat接到请求后出现失败状况或者在Tomcat发送http头给客户端后出现失败状况，则不进行恢复。此属性在jk 1.2.6版本被增加进来，以求避免Tomcat的死机和在支持ajp13的servlet引擎上发生的问题。此属性默认为全部恢复。
因在默认的情况下，JK检测tomcat的工作情况，所以浪费了很多资源，这个资源多到比tomcat正常工作所需的资源的1.6倍还多。如果不检测恢复，则效率高了，处理的请求数可以是原来的2.6倍多，而且tomcat还不死--没响应,反而使出错的数量比检测恢复之后出错的数量还少，性能还加强了，如果原来用这种的集群78台机器，有一阵就挂掉，改成这个配置只要30台机器，一般在同样的时间段内运行无反应或出错的情况比recovery_options使用默认值的少。
recovery_options的最新参考
http://tomcat.apache.org/connectors-doc/reference/workers.html

少林功夫好

回复 9# zawecha1


    虽然是楼主2006年的疑问，等到2010年还有您的回复。真心感谢你的帮忙。撇开问题不管，至少知道您是真心想帮助大家的。再次感谢。