Red Hat Bugzilla – Bug 867467
httpd restart fails for 2 consecutive restarts
Last modified: 2012-10-17 11:18:56 EDT
Description of problem:
When IPA replica install process does 2 consecutive httpd restarts, the second httpd restart may fail and thus IPA replica installation will fail.
See Bug 845405 for details.
Version-Release number of selected component (if applicable):
I think I managed to reproduce it with this simple sequence:
# service httpd restart && service httpd restart
Stopping httpd: [ OK ]
Starting httpd: [ OK ]
Stopping httpd: [FAILED]
Starting httpd: (98)Address already in use: make_sock: could not bind to address [::]:80
(98)Address already in use: make_sock: could not bind to address 0.0.0.0:80
no listening sockets available, shutting down
Unable to open logs
When I tested this sequence on Fedora 18, it worked fine. It just probably waited until the ports could be bound.
Steps to Reproduce:
1. See above - run 2 consecutive service restarts
Second restart fails.
Second restart succeeds.
The answer is "Don't do that", I'm afraid.
With the sysV init scripts we can't make a strong guarantee that httpd is immediately in a state where the service script will always DTRT. Adding a sleep(1) or similar should help, but if you are loading weird modules like mod_nss with complex startup processing, you might need more than 1 second delay.
With systemd we do make a strong guarantee and this kind of thing really works properly.