Apache Dying Weekly
by damonp on June 1, 2005
in SysAdmin
Has anyone had the problem with Apache dying after being up over a week with the following message?
[Wed Jun 01 18:25:01 2005] [crit] (28)No space left on device: mod_rewrite: could not create rewrite_log_lock
There is a moderate amount of discussion on it, but no solutions I found. Memory leaks.
The temporary solution to clear up the semaphores for Apache with a:
ipcs -s | grep nobody | perl -e 'while (<stdin>) { @a=split(/\s+/); print `ipcrm sem $a[1]`}'
ipcs -m | grep nobody | perl -e 'while (</stdin><stdin>) { @a=split(/\s+/); print `ipcrm -m $a[1]`}' </stdin>
works but, it only treats the symptom (dead Apache) not the cause. SIM and cleaning the semaphores daily is the best I have found.
Popularity: 1%
Most Popular Posts

Damon Parker is a freelance sysadmin and web developer in Texas. He specializes
in server setup, server security and high performance server configurations.
Need help setting up a web server or getting a server back online after a crash or
hack? Email Damon
{ 1 trackback }