One of our Linux servers crashes all of a sudden

Hi,

One of our Linux servers crashes all of a sudden rather frequently
That is all JSP based pages will stop loading.
At which point I have tried:
1- restarting Apache
2- Restarting MySQL
3- Restarting Jakarta tomcat

But nothing will recover from the crash but a reboot.

Also I have started noticing this strange message on the Putty window:
Unknown HZ value! (90) Assume 100.

Appreciate your help.

Regards,

Hello,

Well 1st what is strange about this HZ value is that it just started showing up a short while ago, wasnt there before then. And it seems about the same time we started having the JSP crashes, that is JSP based pages will not load which causes relative crash of the site since about 50% of the functions of the site are in JSP. Damn those JSP developers :slight_smile:

About the server OS, it is not custom compiled.
It is Red hat enterprise.
And about dmesg, No I dont see anything strange. In fact I can send it to you via private message so you can see it.

Regards,
Dean.

For the error logs, try looking for files called:
PHP: php_error.log
MySQL: mysql.log
Apache: apache_error.log

Does the php run ok with jsp disabled?

Hi,

Where do I find error log files for MySQL? Apache?
Also the whole server is not crashing, just the Jsp related pages stop loading.
Everything else, in particular Php pages are loading fine.

Regards,

Hi,

I used the Linux locate command to look for these files, but it cannot find any of them!
Example:

locate apache_error.log
locate: warning: database /var/lib/slocate/slocate.db’ is more than 8 days old

[5]+ Stopped locate apache_error.log

And to your other question:
Yes, Php does run when these crashes happen, which means when all Jsp based pages or Jsp based codes will stop loading or running.

Regards,

Hi,

Off the top of my head HZ value is what is called the “tick” value, it’s basically to do with how many times per second the kernel is interrupted to allow data processing etc. 90 is an odd value for it to be set at, usually it’s be 100, 250 or 1000. How old is the server? Is it a custom compiled kernel or a stock OS distro one? Also is there anything odd when you run: dmesg or check /var/log/messages

Thanks,

Is there any entries dated around times when the server is crashing in the error logs for MySQL, Apache and wahever server-side language that stand out as unusual? Any unusual entries in any FTP logs?