Hi.

I found on my hosting a very unpleasant thing, which leads to the
appearance of 502's errors on the users' site.

Here's how it looks from the TCP side:

proxy_balance       <->       peruser
             -> SYN
                   SYN/ACK <-
             -> ACK
             -> REQUEST (GET/POST)
                       ACK <-
                       (wait about 2 seconds)
                       FIN <-
             -> RESET
<- 502 ERROR

It means that the client sends a request to the peruser, it accepts it,
waits for about two seconds and then closes the connection.

Here is what I obtained from the log file. This is an example of a
normal session:
...
[Thu Jul 14 18:06:23 2011] [warn] (peruser: pid=15495 uid=364813 child=4664) 
peruser_post_read(): function entered
[Thu Jul 14 18:06:23 2011] [warn] (peruser: pid=15495 uid=364813 child=4664) 
peruser_post_read(): PROCESSOR 4664
[Thu Jul 14 18:06:23 2011] [warn] (peruser: pid=15495 uid=364813 child=4664) 
peruser_post_read(): request for 8tuv.ru / (server www.8tuv.ru) seems to be for 
us
[Thu Jul 14 18:06:23 2011] [warn] (peruser: pid=15495 uid=364813 child=4664) 
peruser_post_read(): performing handle cleanup
[Thu Jul 14 18:06:23 2011] [warn] (peruser: pid=15495 uid=364813 child=4664) 
child_main(): returned from process_socket()
[Thu Jul 14 18:06:23 2011] [warn] (peruser: pid=15495 uid=364813 child=4664) 
child_main(): Child 4664 (PROCESSOR) is now ready
[Thu Jul 14 18:06:34 2011] [warn] (peruser: pid=15495 uid=364813 child=4664) 
just_die(): function called
[Thu Jul 14 18:06:34 2011] [warn] (peruser: pid=15495 uid=364813 child=4664) 
clean_child_exit(): close(CHILD_INFO_TABLE[4664].senv->input) = 0
[Thu Jul 14 18:06:34 2011] [warn] (peruser: pid=15495 uid=364813 child=4664) 
clean_child_exit(): close(CHILD_INFO_TABLE[4664].senv->output) = 0
[Thu Jul 14 18:06:36 2011] [warn] (peruser: pid=10796 uid=0 child=0) 
ap_mpm_run(): child #4664 (pid=15495) has died

and this with  the premature connection closing:
...
[Thu Jul 14 19:10:02 2011] [warn] (peruser: pid=24249 uid=372056 child=3863) 
peruser_process_connection(): function entered
[Thu Jul 14 19:10:02 2011] [warn] (peruser: pid=24249 uid=372056 child=3863) 
peruser_post_read(): function entered
[Thu Jul 14 19:10:02 2011] [warn] (peruser: pid=24249 uid=372056 child=3863) 
peruser_post_read(): PROCESSOR 3863
[Thu Jul 14 19:10:02 2011] [warn] (peruser: pid=24249 uid=372056 child=3863) 
peruser_post_read(): request for 1eka.ru / (server www.1kra.ru) seems to be for 
us
[Thu Jul 14 19:10:02 2011] [warn] (peruser: pid=24249 uid=372056 child=3863) 
peruser_post_read(): performing handle cleanup
[Thu Jul 14 19:10:05 2011] [warn] (peruser: pid=7113 uid=0 child=0) 
ap_mpm_run(): child #3863 (pid=24249) has died

And there're plenty of such errors.
What could it be?

Thank you in advance.

-- 
BRGDS. Alexey Vlasov.
_______________________________________________
Peruser mailing list
[email protected]
http://www.telana.com/mailman/listinfo/peruser

Reply via email to