>> mod_jk is not the speciality of Pier :)
>
>Sorry ;-)) My eyes are not so good ;-)

:::)

>> >I had (on linux rh 6 and 7).
>>
>> What's the exact problem with mod_jk and ajp12 ?
>> Couldn't you switch to tomcat 3.2.3 and ajp13 ?

Exactly the same configuration here in dev and prod.
Redhat 6.x, TC 3.2.3/3.3 and mod_jk / ajp13
Never go any problems

>Never thought about that option actually. We used the default 
>ajp (in our
>case indeed ajp12) which was with the distribution. (from the 
>conf-auto).
>Let's say our experience was horrible with it ;-) I can remember a big
>discussion on tomcat-user however with a lot of people having the same
>problems (I unsubscribed at a certain moment, so I lost if it ever got
>fixed).

Problems with ajp13 ? It's the reference today and all the effort are
against mod_jk/ajp13 duo. I don't see bugs using ajp13.

>What happens after a certain amount of time is that 
>libhttpd.ep is at 50%
>(times 4, we have 2 cpu's in the production server..). It still served
>requests, but It was like having my Sinclair ZX 81 running windows NT
>-(( (that is really bad..). It never "crashed" when people 
>were using the
>shops created with our toolkit (just a little session stuff), 
>but happened
>when people were changing or creating shops (my guess is the 
>upload forms
>that would mess up the system..).

What's libhttpd.ep ? Don't you said you're running Linux Redhat ? In that
case process name is httpd.

>Searched for some old mod_jk errors and here they are some entries.. :
>
>[jk_ajp12_worker.c (596)]: ajpv12_handle_response, error 
>writing back to
>server (and then a lot of times..)
>and some of these :
>[jk_ajp12_worker.c (463)]: In ajpv12_handle_request, failed to 
>write to the
>cont
>ainer
>[jk_ajp12_worker.c (152)]: In jk_endpoint_t::service, Error sd = 6

Did you restart tomcat at some time ?

>If you want I can zip up our logfile and send it to your 
>private mail ;-))

Yes you could but ajp12 IS CLEARLY DEPRECATED today. Only used to
shutdown Tomcat.

Reply via email to