Started off by

  1. Enabling unattended updates
  2. Enable only ssh login with key
  3. Create user with sudo privileges
  4. Disable root login
  5. Enable ufw with necessary ports
  6. Disable ping
  7. Change ssh default port 21 to something else.

Got the ideas from networkchuck

Did this on the proxmox host as well as all VMs.

Any suggestions?

  • theniwo@alien.topBEnglish
    1·
    1 year ago

    Enabling unattended updates -> Hell no. Regular Patchdays
    Enable only ssh login with key -> yes
    Create user with sudo privileges -> yes
    Disable root login -> no
    Enable ufw with necessary ports -> Basic iptables, but not on all hosts. But fail2ban
    Disable ping -> nope
    Change ssh default port 21 to something else. -> nope

      • theniwo@alien.topBEnglish
        1·
        1 year ago

        Can you give me ressources on how to configure f2b?

        I usually leave the defaults, or maybe tweak the times a bit.

        One could only enter my network thru vpn or nginx on 443 anyway, so I am not that worried

        • lack_of_reserves@alien.topBEnglish
          1·
          1 year ago

          The majority of the default fail2ban installations only bans an IP for 10 minutes and uses a 10 minute findtime, e.g. slow brute forcing is not at all banned.

          Before I switched to crowdsec (which I really recommend you do, its quite easy) I changed my bantime and findtime in /etc/fail2ban/jail.conf (I think I made a local file… read the file it should say) to something like 8 hours (e.g. change 10m to 640m for both those variables).

          • theniwo@alien.topBEnglish
            1·
            1 year ago

            Well if you are using strong passwords or no passwords from outside at all, but key auth only, i think you are pretty in the safe side. As i said, i have no ssh port open to the internet. Raising the ban time could only lead to banning myself. 😀

            But for ports open to the outside, yes. I ppbly would do that too. Plus hardening the ssh config a bit

            • lack_of_reserves@alien.topBEnglish
              1·
              1 year ago

              I have an open ssh port and I use key auth with password as well as crowdsec. Even if people get my ssh key they would still need to know the password.