Welcome to WebmasterWorld Guest from 54.210.61.41

Forum Moderators: Ocean10000 & phranque

Message Too Old, No Replies

can mod_gzip cause apache to crash?

after it has been running for some time

     
5:08 pm on Mar 30, 2003 (gmt 0)

Senior Member from ES 

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month

joined:July 24, 2002
posts:1128
votes: 2


hi,

i installed the newest mod_gzip on to our redhat 7.2 server last night.

everything worked perfectly - was lightning quick.

however today about 3 hours ago our server went offline. all domains were inaccessible, SSH and FTP were inaccessible. but i was able to ping the server normally.

at the moment the server is being rebooted (that's when the hosts gets around to it, grrrr)

they suggested it might be a CPU process.

can anyone give me some tips on how to go about identifying what caused apache (at least i think it must be apache) to crash, as I am a relative beginner.

can it be mod_gzip, even though it appeared to work fine for 18 hours or so.

much appreciated!

p.s. we are running a P4 2Ghz and are only serving 50.000 pages a day, so it can't be cpu resources. furthermore by coincidence i was looking at TOP this afternoon before the crash and CPU usage was 0.1%.

7:39 pm on Mar 30, 2003 (gmt 0)

Senior Member from ES 

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month

joined:July 24, 2002
posts:1128
votes: 2


now back online, this is from /var/log/messages

Mar 30 14:51:13 eul0000533 kernel: Out of Memory: Killed process 26341 (httpd).
Mar 30 14:51:13 eul0000533 kernel: Out of Memory: Killed process 26333 (httpd).
Mar 30 14:51:39 eul0000533 kernel: Out of Memory: Killed process 27030 (httpd).
and many more

and this from error_log

[Sun Mar 30 14:04:26 2003] [notice] SIGUSR1 received. Doing graceful restart
this is me restarting apache earlier today after fiddling around
[Sun Mar 30 14:04:27 2003] [notice] Apache/1.3.27 (Unix) PHP/4.1.2 mod_gzip/1.3.26.1a configured -- resuming normal operations
[Sun Mar 30 14:04:27 2003] [notice] suEXEC mechanism enabled (wrapper: /usr/local/apache/bin/suexec)
[Sun Mar 30 14:04:27 2003] [notice] Accept mutex: sysvsem (Default: sysvsem)
[Sun Mar 30 14:53:26 2003] [notice] child pid 27170 exit signal Alarm clock (14)
the following is after rebooting
[Sun Mar 30 17:28:59 2003] [warn] pid file /usr/local/apache/logs/httpd.pid overwritten -- Unclean shutdown of previous Apache run?

does anyone have any suggestions? could this be a mod_gzip problem, or should i look elsewhere?

many thanks

p.s. as a precaution i have excluded php files (our forum is busy) from gzipping - not sure if that will help or not

8:55 am on Mar 31, 2003 (gmt 0)

Senior Member from ES 

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month

joined:July 24, 2002
posts:1128
votes: 2


i think i have identified one of possible problems:

i had loaded the modules in the wrong order. php was loaded after gzip in httpd.conf. i have now changed the order and it reads correctly:

[Mon Mar 31 08:48:31 2003] [notice] Apache/1.3.27 (Unix) mod_gzip/1.3.26.1a PHP/4.1.2 configured -- resuming normal operations

i don't know whether this could have caused the crash. and since then i have excluded .php files from compression, so this will probably have no effect.

hmmm...

i have also been monitoring cpu% via top command and i noticed no difference between having gzip turned on or off. i am fairly confident that it is not gzip which caused the crash, or at least am confident that the way it is configured now shouldn't cause future problems. this pleases me, as the speed with gzip is incredible. would however still like to identify the real reason for the crash.

anyway enough ramblings

 

Join The Conversation

Moderators and Top Contributors

Hot Threads This Week

Featured Threads

Free SEO Tools

Hire Expert Members