after upgrade, httpd_ls_bak error

[aww]

I've always had this httpd_ls_bak error after upgrades, seems minor but figure I better report it:

Would you like to restart it right now [Y/n]?

[OK] Lease License key #2 will expire in XX days!
/usr/local/apache/bin/httpd_ls_bak: invalid option -- k
Usage: /usr/local/apache/bin/httpd_ls_bak [-D name] [-d directory] [-f file]
[-C "directive"] [-c "directive"]
[-v] [-V] [-h] [-l] [-L] [-G] [-S] [-t] [-T] [-F]
Options:

 

[mistwang]

The error message is mainly because some version of Apache support "-k" option, some version does not. We are aware of it and will address it later.

Thanks!

 

[iCARus]

We got on one server similar problems. LSWS just stop working. Proces lshttpd hits CPU to 100% for couple of minutes. We tried to restart/stop/start lsws and is just ignoring.

After couple of starts working again.

strace.log

select(1, [0], NULL, NULL, {1, 0}) = 0 (Timeout)
kill(8715, SIG_0) = 0
select(1, [0], NULL, NULL, {1, 0}) = 0 (Timeout)
kill(8715, SIG_0) = 0
select(1, [0], NULL, NULL, {1, 0}) = 0 (Timeout)
kill(8715, SIG_0) = 0
select(1, [0], NULL, NULL, {1, 0}) = 0 (Timeout)
kill(8715, SIG_0) = 0
select(1, [0], NULL, NULL, {1, 0}) = 0 (Timeout)
kill(8715, SIG_0) = 0
select(1, [0], NULL, NULL, {1, 0}) = 0 (Timeout)
kill(8715, SIG_0) = 0
select(1, [0], NULL, NULL, {1, 0}) = 0 (Timeout)
kill(8715, SIG_0) = 0
select(1, [0], NULL, NULL, {1, 0}) = 0 (Timeout)
kill(8715, SIG_0) = 0
select(1, [0], NULL, NULL, {1, 0}) = 0 (Timeout)
kill(8715, SIG_0) = 0
select(1, [0], NULL, NULL, {1, 0}) = 0 (Timeout)
kill(8715, SIG_0) = 0
select(1, [0], NULL, NULL, {1, 0}) = 0 (Timeout)
kill(8715, SIG_0) = 0
select(1, [0], NULL, NULL, {1, 0}) = 0 (Timeout)
kill(8715, SIG_0) = 0
select(1, [0], NULL, NULL, {1, 0}) = 0 (Timeout)
kill(8715, SIG_0) = 0
select(1, [0], NULL, NULL, {1, 0}) = ? ERESTARTNOHAND (To be restarted)
--- SIGINT (Interrupt) @ 0 (0) ---
+++ killed by SIGINT +++

error.log

2008-07-27 10:42:05.976 [ERROR] HttpListener::start(): Can't listen at address adminListener: Address already in use!
2008-07-27 10:42:05.976 [ERROR] HttpServer::addListener(adminListener) failed to create new listener
2008-07-27 10:42:05.976 [ERROR] [config:admin:listener:adminListener] failed to start listener on address *:7080!
2008-07-27 10:42:05.976 [ERROR] [config:admin:listener] No listener is available for admin virtual host!
2008-07-27 10:42:05.976 [ERROR] Fatal error in configuration, exit!

This LSWS works before upgrade great without any downtimes. Now every day we suffer this strange bug/problem.

Any idea ?

P.S.
we are now swiching back to v3.3.14 to see if this happen again.

 

[mistwang]

We have noticed the "failed to start listener on address *:7080" problem on one server, we will have it addressed in 3.3.16.

But the 100% cpu of lshttpd is something else, if you notice that again, please strace it, then "kill -USR2 <pid_of_100%_CPU_lshttpd>" to turn on debug logging, let it run for a little while, then kill it. send us the error log and strace log.

 

[iCARus]

This problem exist in v3.3.14 too.

But strange:

After lshttpd is on 100% CPU, is like this

2031 nobody 15 0 29612 23m 8880 S 100 0.6 0:01.41 lshttpd

If everything is ok then top shows(with slash and dot):

2031 nobody 15 0 29612 23m 8880 S 1 0.6 0:01.41 ./lshttpd

This is really annoying while every day we go downtime of couple of min.

 

[mistwang]

The problem is mainly caused by high I/O wait due to more than 20,000 core files under /usr/local/apache/htdoc/mrtg/ directory. After removing those core files, the server run pretty smooth.

If notice more problem just let me know.

 

[iCARus]

This problem still exist. We don't have anymore MRTG installed.