For abuse issues that involve your server host sending emails with complaints that your server is conducting network scanning.

What is Network Scanning?

Network scanning is a process of identifying active hosts on a network, either for the purpose of attacking them or for network security assessment. It laymans terms, if your hosting provider has sent you an email then your system is compromised and has a script on it that is scanning other systems.

The tool that can detect outgoing portscans is tcpdump.

Install tcdump:

yum install tcpdump -y

Run

# tcpdump -i eth0 -w dump 

and then, with a lot of calm, read the dump details matching the packets sent in those dates.

Ask your users for their IPs also, if static.

Look to /var/log/auth.log and last -100 to see if someone abused your system, check for suspicious cron jobs, look into the /root directory if something strange appeared. In these cases a tool that hashize the system files (like tripwire) would be important.

It is recommended that a virus or rootkit tool be used to scan the websites on the server for any shell scripts or exploits that may have allowed for outbound SSH attacks.

We have an article on installing and using ClamAV ( http://geekdecoder.com/clamav-on-centos-6/ ).

For plesk – I would recommend running on /var/www/vhosts (e.g. clamscan -ir /var/www/vhosts/ –log=/root/clamscan-01.07.2015.txt).
For cpanel – I would recommend running on /home (e.g. clamscan -ir /home –log=/root/clamscan-01.07.2015.txt).

Rkhunter is another scanning tool that may identify any uploaded malicious files ( http://www.tecmint.com/install-linux-rkhunter-rootkit-hunter-in-rhel-centos-and-fedora/ ).

Once any found shells are removed, if it were within /var/www/vhosts for plesk or /home for cpanel, it’s likely that the exploit was through one of the sites. If the file was found within a site’s document root, I would update any licensed or open source software to remove potential vulnerabilities.

Also, to disable any outgoing SSH usage if the scans were ssh scans(and if you do not use SSH from the server, it resolves the symptom immediately), a firewall rule can be added (iptables -I OUTPUT -p tcp –dport 22 -j DROP).