Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Our server is forced to restart sometimes because some users (not all) could start more zombie suphp defunct processes than allowed through the RLIMIT_NPROC / ulimit -u statement.

How is that possible? Is there a known bug or something similar? And do you know how I could handle that?

I'm running CentOS 5.8 with Apache 2.2.3

Thanks. Best regards, John

share|improve this question
    
What process is the parent of these defunct processes? – David Schwartz Mar 20 '13 at 13:59
    
Hm, good question. Could I figure that out backwards? Where is this logged? – John Brunner Mar 20 '13 at 14:29
    
don't really know which parent processes our users run, but this rlimit_nproc statement is declared for every user (in apache config), in the range of 20-50, and some user could exceed that limit, and other user have troubles to run under 10 processes. it's not really reliable – John Brunner Mar 20 '13 at 14:41
    
I've made an c script which produces 50,000 zombie processes. even if my ulimit -u statement is configured to 50,000... there are just exactly 32,354 processes (not more). when I lower down the ulimit it works finde (ulimit -> 20,000 zombie processes -> 20,001)... i can't figure out why it works sometimes and sometimes not – John Brunner Mar 20 '13 at 15:49

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.