0

I continue to get HTTPD failure messages from WHM sent to me, seemingly random... cPanel ChkServd Service Monitor httpd on vds64_centos54 failed. httpd failed A restart was attempted automagically. Failure reason: Service check failed to complete httpd: [!= http/1...] I'm at restart attempt number 18 now...

here is the result of this:

[root@server ~]# service httpd status

Looking up localhost
Making HTTP connection to localhost
Sending HTTP request.
HTTP request sent; waiting for response.
Alert!: Unexpected network read error; connection aborted.
Can't Access `http://localhost/whm-server-status'
Alert!: Unable to access document.

lynx: Can't access startfile

Everything seems to be running normal on the server though to the public...

Some other info... This is from /usr/local/apache/logs/error_logs:

[Sat Aug 20 16:03:54 2011] [notice] child pid 12852 exit signal Segmentation fault (11)
[Sat Aug 20 16:04:07 2011] [notice] child pid 12854 exit signal Segmentation fault (11)
[Sat Aug 20 16:05:02 2011] [notice] child pid 12857 exit signal Segmentation fault (11)
[Sat Aug 20 16:05:21 2011] [notice] child pid 12850 exit signal Segmentation fault (11)
[Sat Aug 20 16:07:54 2011] [notice] caught SIGTERM, shutting down

It looks as if I get the message from CPanel anytime I see the CAUGHT SIGTERM message

Shackrock
  • 198
  • 5
  • 16
  • What does the webserver log say when you try that command? – DerfK Aug 20 '11 at 20:08
  • @derfK Check my update. Which logs are you looking for, what is the path? – Shackrock Aug 20 '11 at 20:15
  • SIGTERM is from shutting down or restarting the server. It's possible that the server is being restarted because it fails to get the status. If you can, in one window/screen run `tail -f /usr/local/apache/logs/error_logs` to watch the log as it's being written, and in the other run `service httpd status`. Is that the cause of the Segmentation fault message? – DerfK Aug 20 '11 at 20:36
  • @DerfK Looks like running that command creates: `[Sat Aug 20 16:39:46 2011] [notice] child pid 14369 exit signal Segmentation fault (11)` to come into the log. The restart comes every 5 minutes like clockwork... But again, the server is running fine to the public... – Shackrock Aug 20 '11 at 20:41
  • @DerfK I should add... I think this started to happen either after installing Atomic, or possibly from trying to get mcrypt package installed. I don't really want/need Atomic - was just checking it out. How can I uninstall that without messing anything up? I also still have issues with mcrypt: http://serverfault.com/questions/303329/installing-php-mcrypt-library-libmcrypt-on-centos-5-server-help-please/303379 – Shackrock Aug 20 '11 at 20:43
  • 1
    see edit on my answer – DerfK Aug 20 '11 at 21:23
  • @DerfK let us [continue this discussion in chat](http://chat.stackexchange.com/rooms/1136/discussion-between-shackrock-and-derfk) – Shackrock Aug 20 '11 at 21:32

1 Answers1

2

The fact that requesting the webserver status is causing the server to segfault implies that your mod_status module is broken in some way. Check apache's configuration to make sure it is not loading a stray mod_status module from some older installation that doesn't match the current server.

EDIT

OK, based on the other question it looks like your webserver was messed up by that failed httpd.x86_64 0:2.2.3-45.el5.centos.1 package. It's possible that it overwrote some things before it errored. Were you using the EasyApache package before? You will probably have to rebuild it to get things working right again. If you go this route you'll have to use EasyApache to customize the php modules that are installed rather than using centos packages. If you use the centos packages, you may have to manually reconfigure the server to suport the cpanel/whm stuff (like the custom whm-server-status link). From your log in the other question, it appears that the old httpd configuration was not overwritten, so you may be able to merge together the configuration files.

EDIT 2

Based on the chat here, it appears that what has happened is that installing Atomic Secured Linux overwrote the mod_security2.so module installed by EasyApache with an incompatible version, and this somehow only trigged the segfault when accessing the status page. Disabling the incompatible mod_security2.so fixes the issue, but you'll need to rebuild EasyApache in order to get back the mod_security2.so that was overwritten, and how to do that (or if it's possible to just rebuild the module) is a matter for a new question.

DerfK
  • 19,313
  • 2
  • 35
  • 51
  • hmm, that makes sense - maybe it's just a bunk status check - gotcha. So, can you help direct me on how to check that..? Sorry, totally new to me... – Shackrock Aug 20 '11 at 21:01