apache getting stuck
- Thread starter Wunk
- Start date
Just replace the restart part of /etc/init.d/httpd with:
Works for me
Code:
restart)
stop
waitforexit "httpd" 20
kill -9 `netstat -lnp | grep ':80 ' | awk '{print $7}' | awk -F "/" '{print $1}'`
kill -9 `netstat -lnp | grep ':443 ' | awk '{print $7}' | awk -F "/" '{print $1}'`
start
;;Works for me
vtx, yes, but it's only a workaround.
I know, but better a workaround than an apache getting stuck at random nights
Just replace the restart part of /etc/init.d/httpd with:
Code:restart) stop waitforexit "httpd" 20 kill -9 `netstat -lnp | grep ':80 ' | awk '{print $7}' | awk -F "/" '{print $1}'` kill -9 `netstat -lnp | grep ':443 ' | awk '{print $7}' | awk -F "/" '{print $1}'` start ;;
Works for me
this works, thank you
nobaloney
NoBaloney Internet Svcs - In Memoriam †
At least several posters have had the problem and it wasn't exim, so I'm not sure what this proves.
Martynas, have you ever seen exim.pl in the netstat output? exim.pl of course isn't a daemon; it should be called for each email but of course shouldn't hang around.
John, what do you think? Is there anything in exim.pl that you think could be causing the problem?
Jeff
Martynas, have you ever seen exim.pl in the netstat output? exim.pl of course isn't a daemon; it should be called for each email but of course shouldn't hang around.
John, what do you think? Is there anything in exim.pl that you think could be causing the problem?
Jeff
Jeff, yes, I'm sure exim is causing the problem. Exim.pl isn't a daemon, but it's used by exim daemon. Here you are lsof As you see - it's using apache's files.
As you see - it's using apache's files.
Code:
exim 15135 root 7w REG 8,2 8240
29016088 /var/log/httpd/error_log
exim 15135 root 8w REG 8,2 1660
511063 /var/log/httpd/domains/dddd.info.error.log
exim 15135 root 9w REG 8,2 481770
511064 /var/log/httpd/domains/ccccc.com.error.log
exim 15135 root 10w REG 8,2 561481
511065 /var/log/httpd/domains/bbbbb.com.share.error.log
exim 15135 root 11w REG 8,2 6013
511066 /var/log/httpd/domains/bbbbb.com.error.log
exim 15135 root 12w REG 8,2 461928
511067 /var/log/httpd/domains/dddd.in.error.log
exim 15135 root 13w REG 8,2 1386
511068 /var/log/httpd/domains/aaaa.com.error.log
exim 15135 root 14w REG 8,2 12651
29016526 /var/log/httpd/access_log
exim 15135 root 15w REG 8,2 12
511061 /var/log/httpd/domains/dddd.info.bytes
exim 15135 root 16w REG 8,2 739
511069 /var/log/httpd/domains/dddd.info.log
exim 15135 root 17w REG 8,2 12
511061 /var/log/httpd/domains/dddd.info.bytes
exim 15135 root 18w REG 8,2 739
511069 /var/log/httpd/domains/dddd.info.log
exim 15135 root 19w REG 8,2 752054
511057 /var/log/httpd/domains/ccccc.com.bytes
exim 15135 root 20w REG 8,2 109257281
511070 /var/log/httpd/domains/ccccc.com.log
exim 15135 root 21w REG 8,2 752054
511057 /var/log/httpd/domains/ccccc.com.bytes
exim 15135 root 22w REG 8,2 109257281
511070 /var/log/httpd/domains/ccccc.com.log
exim 15135 root 23w REG 8,2 76224
511040 /var/log/httpd/domains/bbbbb.com.share.bytes
exim 15135 root 24w REG 8,2 3734211
511071 /var/log/httpd/domains/bbbbb.com.share.log
exim 15135 root 25w REG 8,2 76224
511040 /var/log/httpd/domains/bbbbb.com.share.bytes
exim 15135 root 26w REG 8,2 3734211
511071 /var/log/httpd/domains/bbbbb.com.share.log
exim 15135 root 27w REG 8,2 4881
511052 /var/log/httpd/domains/bbbbb.com.bytes
exim 15135 root 28w REG 8,2 192146
511072 /var/log/httpd/domains/bbbbb.com.log
exim 15135 root 29w REG 8,2 4881
511052 /var/log/httpd/domains/bbbbb.com.bytes
exim 15135 root 30w REG 8,2 192146
511072 /var/log/httpd/domains/bbbbb.com.log
exim 15135 root 31w REG 8,2 797992
511045 /var/log/httpd/domains/dddd.in.bytes
exim 15135 root 32w REG 8,2 56896838
511073 /var/log/httpd/domains/dddd.in.log
exim 15135 root 33w REG 8,2 797992
511045 /var/log/httpd/domains/dddd.in.bytes
exim 15135 root 34w REG 8,2 56896838
511073 /var/log/httpd/domains/dddd.in.log
exim 15135 root 35w REG 8,2 0
511049 /var/log/httpd/domains/aaaa.com.bytes
exim 15135 root 36w REG 8,2 0
511074 /var/log/httpd/domains/aaaa.com.log
exim 15135 root 37w REG 8,2 0
511049 /var/log/httpd/domains/aaaa.com.bytes
exim 15135 root 38w REG 8,2 0
511074 /var/log/httpd/domains/aaaa.com.log
exim 15135 root 39w REG 8,2 73264
29016529 /var/log/httpd/homedir.log
exim 15135 root 40w REG 8,2 73264
29016529 /var/log/httpd/homedir.log
exim 15135 root 41w REG 8,2 73264
29016529 /var/log/httpd/homedir.log
exim 15135 root 42w REG 8,2 73264
29016529 /var/log/httpd/homedir.log
exim 15135 root 43w REG 8,2 73264
29016529 /var/log/httpd/homedir.log
exim 15135 root 44w REG 8,2 73264
29016529 /var/log/httpd/homedir.log
exim 15135 root 45w REG 8,2 73264
29016529 /var/log/httpd/homedir.log
exim 15135 root 46w REG 8,2 73264
29016529 /var/log/httpd/homedir.log
exim 15135 root 47w REG 8,2 12651
29016526 /var/log/httpd/access_log
exim 15135 root 48w REG 8,2 0
29016530 /var/log/httpd/ssl_request_log
exim 15135 root 49w REG 8,2 0
29016882 /var/log/httpd/ssl_mutex.14811 (deleted)
exim 15135 root 3u IPv4 157531 TCP *:http
(LISTEN)
exim 15135 root 4u IPv4 157536 TCP
*:https (LISTEN)
exim 15136 mail 3u IPv4 157531 TCP *:http
(LISTEN)
exim 15136 mail 4u IPv4 157536 TCP
*:https (LISTEN)
Last edited:
Ah.. and before lsof, "service apache restart" was done with this result:
So, only "killall -9 exim" could help in this situation
Code:
Stopping httpd: [FAILED]
Starting httpd: (98)Address already in use: make_sock: could not bind to
address 0.0.0.0:443
no listening sockets available, shutting down
Unable to open logs
[FAILED]So, only "killall -9 exim" could help in this situation
nobaloney
NoBaloney Internet Svcs - In Memoriam †
If it's exim.pl, couldn't we just do:
?
Should one of these killall lines be added to the httpd restart command? Or even to the httpd start command?
Jeff
Code:
killall -9 exim.plShould one of these killall lines be added to the httpd restart command? Or even to the httpd start command?
Jeff
exim.pl is used by /usr/sbin/exim (and exim.pl isn't a process), because of
So, it's "included" into exim We shouldn't use workarounds, we should use real fixes, don't you think so?
Code:
perl_startup = do '/etc/exim.pl'So, it's "included" into exim
We shouldn't use workarounds, we should use real fixes, don't you think so?
nobaloney
NoBaloney Internet Svcs - In Memoriam †
Sure. Create the fix !
That's the beauty of opensource, you can do it instead of complain about it
.
Jeff
!That's the beauty of opensource, you can do it instead of complain about it
.Jeff
selfwebhosting
Verified User
Sure. Create the fix
That's the beauty of opensource, you can do it instead of complain about it
Jeff
Has anyone created a fix?
Before a fix is found, I have to use a cronjob to restart apache every four hours a day!
I am running apache2 and I have commented out the ErrorLog lines in all the virtual_host*.conf template config files.
nobaloney
NoBaloney Internet Svcs - In Memoriam †
exim.pl IS found in the process list when it's running. And that change I mentioned to the apache script may very well resolve the problem. If it does, it IS the fix, and adding it to the restart script IS the right way to fix it.
But someone needs to test it first.
Jeff
But someone needs to test it first.
Jeff
selfwebhosting
Verified User
In my case, it is both exim and MySQL. So how should I modify the restart script? Yes, I am willing to test it.
jlasman, as you see in lsof - it's exim, and not exim.pl, but in exim (not exim.pl) there are no such paths as /var/log/httpd/domains/aaaa.com.bytes etc. Don't you think so? "killall -9" won't ever be a fix, because it IS really a workaround, I never recommend to use "-9" flag, because it can damage the process (file or something more), if you don't think that it's true - try to kill mysql with "-9" few times a day and you will see some damaged databases.
selfwebhosting
Verified User
Jeff, yes, I'm sure exim is causing the problem. Exim.pl isn't a daemon, but it's used by exim daemon. Here you are lsof
Code:exim 15135 root 7w REG 8,2 8240 29016088 /var/log/httpd/error_log exim 15135 root 8w REG 8,2 1660 511063 /var/log/httpd/domains/dddd.info.error.log exim 15135 root 9w REG 8,2 481770 511064 /var/log/httpd/domains/ccccc.com.error.log ... exim 15135 root 49w REG 8,2 0 29016882 /var/log/httpd/ssl_mutex.14811 (deleted) exim 15135 root 3u IPv4 157531 TCP *:http (LISTEN) exim 15135 root 4u IPv4 157536 TCP *:https (LISTEN) exim 15136 mail 3u IPv4 157531 TCP *:http (LISTEN) exim 15136 mail 4u IPv4 157536 TCP *:https (LISTEN)
smtalk, how did you capture the above?
In my case it is definitely MySQL as when the load is high I see a lot of MySQL connections. After clicking "stop" followed by "start" for MySQL, the connection numbers suddenly drop down to 2 or 3. Also when the CPU upload shoots up to 16, I see MySQL is taking up 99% of the CPU.
How can I pin-point which domain/script is causing this problem? Now I am running vtx's modified script for the apache restart section every two hours. But from the DA control panel I see I can manually "stop" and then "start" MySQL. How can I do it automatically, that is, writing a simple script and then execute it using a cronjob? Like smtalk says, it killing MySQL with "-9" flag a few times a day can damage databases, how can I avoid using that flad. If I do it manually a few times a day, would this damage my databases?
Last edited:
selfwebhosting
Verified User
OK, I quit using vtx's modified apache restart script. I now run this crontab every two hours:
Would this create problem for my databases? How can I do a graceful restart rathen than a sudden stop and start?
Then the still remaining question of how to track down which database has been used to jump up my CPU usage to 99% - any tip will be appreciated.
Code:
15 */2 * * * /sbin/service mysqld restartWould this create problem for my databases? How can I do a graceful restart rathen than a sudden stop and start?
Then the still remaining question of how to track down which database has been used to jump up my CPU usage to 99% - any tip will be appreciated.