check_fail2ban

In the header block it mentions a staus check being implemented which I think is where this came from at line 18 ps_state=$(ps aux |grep "fail2ban.sock" |grep -v grep| wc -l) This seems to be an attempt to check if the fail2ban.sock socket has been created by the nagios-server, except this is looking at processes not sockets or at least it is on my Centos 7.7 installation, it always returns zero so the program aborts further down with the message ++++ Process is not running ++++ Change line 18 to this instead ps_state=$(ss -aux |grep "fail2ban.sock" |grep -v grep| wc -l) Basically use ss instead of ps and put the - before the aux options. Then it returns 1 as expected if the fail2ban-server is running and the script works correctly.

Output of 0.9.x fail2ban is different (added white characters). This patch fix it: --- a/check_fail2ban.sh +++ b/check_fail2ban.sh @@ -105,7 +105,7 @@ fi # ##################################################################### # Run the fail2ban-client, and generate the list of operating jails # ##################################################################### -jail_list=$($fail2ban_client status|grep "list" |cut -f 3 |tr -d ,) +jail_list=$($fail2ban_client status|grep "list" | cut -d : -f 2 | tr -d ,) # ##################################################################### # Cycle through all jails, count blocked IP, and build output strings