Bug 185923 – net-analyzer/fail2ban-0.8.0-r1 failed to start old /tmp/fail2ban.sock

• Product: Gentoo Linux
• Status: RESOLVED DUPLICATE

[Description oxymoron 2007-07-19 22:01:29 UTC]

there is a problem with init script not handling the presence 
of previous /tmp/fail2ban.sock file. It stops for a while and then  there is a message that it failed. Also there is nothing in log file mentioned.
Resolution was to run fail2ban-client start which ends up with message that 
there is probably old /tmp/fail2ban.sock file .....


Reproducible: Always

[Comment 1 Paweł Hajdan, Jr. (RETIRED) 2007-07-26 13:16:55 UTC]

Actually it's quite easy to fix.

fail2ban-client man page:

-x     force execution of the server (remove socket file)

So /etc/init.d/fail2ban script file should be modified like this:

Now is:

start() {
        ebegin "Starting fail2ban"
        ${FAIL2BAN} start &> /dev/null
        eend $? "Failed to start fail2ban"
}

Should be:

start() {
        ebegin "Starting fail2ban"
        ${FAIL2BAN} -x start &> /dev/null
        eend $? "Failed to start fail2ban"
}

[Comment 2 Roeland Douma 2007-08-13 22:41:25 UTC]

Actually it is even simpler.
Just change your /etc/conf.d/fail2ban

[Comment 3 Raphael Marichez (Falco) (RETIRED) 2007-09-13 22:28:31 UTC]

We won't change anything, see bug 168071. This is the expected behaviour. (not a bug, but a feature ;þ )

Changing the init script is *not* the good way at all. See /etc/conf.d/fail2ban as indicated in comment #2

*** This bug has been marked as a duplicate of bug 168071 ***