Help centre

The credentials email arrives at provisioning. Do this before you install anything else.

1. 01.1

   Log in over SSH. Use the IPv4 from the console — it is also your PTR test subject later.

   ssh root@185.42.16.10

2. 01.2

   Create a user with sudo and copy your key over. Root login over password is how servers get pwned at 3 a.m.

   adduser deploy
   usermod -aG sudo deploy
   ssh-copy-id deploy@185.42.16.10

3. 01.3

   Turn on the firewall before you open anything. Default deny, allow SSH only.

   ufw allow OpenSSH
   ufw enable
   ufw status

4. 01.4

   Update once — our images ship current, but a minute of apt beats a month of patching later.

   apt update && apt upgrade -y   # dnf update -y on Alma/Rocky/Fedora

Mail servers and many IRC networks require a PTR record that matches your hostname.

1. 02.1

   In the console, open the server → Networking → Reverse DNS. Enter the fully qualified name, e.g. mail.yourdomain.example.

2. 02.2

   Propagation takes a couple of minutes. Verify from any machine with dig:

   dig -x 185.42.16.10 +short
   # expect: mail.yourdomain.example.

3. 02.3

   Then make sure the forward record agrees — a PTR that no A record points to is a PTR that gets your mail filtered.

   dig +short mail.yourdomain.example
   # expect: 185.42.16.10

Port 25 is closed by default on every new server, because open relays are how the internet catches cold.

1. 03.1

   Open a ticket from the console → Support, subject 'Port 25 request'. Say what the server will send and from which domain.

   Subject: Port 25 request — mail.yourdomain.example
   Body: outbound SMTP for a self-hosted mail server on
         185.42.16.10. SPF and DKIM published. No bulk mail.

2. 03.2

   A human reviews the request — median first response twelve minutes. Approved requests are opened on the network edge, not on your firewall.

3. 03.3

   Verify from the server itself:

   nc -vz smtp-relay.example 25
   # expect: Connection succeeded

Snapshot restore overwrites the disk in place — IPs, hostname and network stay exactly as they were.

1. 04.1

   In the console, open the server → Snapshots. Pick the point in time you trust and press Restore. The server stops briefly, the disk is rewritten, and it boots back.

2. 04.2

   Restores take two to five minutes on Gen4 NVMe. The console shows the job; do not power-cycle while it runs.

3. 04.3

   When it is back, check the boot time and that the disk is the restored size:

   uptime
   df -h /

Extra vCPU, memory, NVMe and snapshots apply live, billed proportionally for the rest of the term.

1. 05.1

   Console → server → Resources. Add what you need; the price shown is the full-term difference, payable once, and the invoice lands in the console.

2. 05.2

   Memory and vCPUs hot-add on Linux with no reboot — confirm with:

   free -h
   nproc

3. 05.3

   Grown NVMe needs one partition step on the guest (the volume itself is already enlarged):

   sudo growpart /dev/vda 1 && sudo resize2fs /dev/vda1

Bad firewall rule, lost key, sshd edited into a coma — rescue mode mounts your disk while the OS stays off.

1. 06.1

   Console → server → Power → Boot into rescue. The server reboots into a network rescue image with a temporary root password shown on screen.

2. 06.2

   Open the VNC console, log in, and mount the disk:

   lsblk                        # find the root partition, e.g. vda2
   mkdir -p /mnt/disk && mount /dev/vda2 /mnt/disk

3. 06.3

   Fix the cause — undo the firewall rule or reset the password:

   chroot /mnt/disk /bin/bash
   ufw allow OpenSSH           # undo whatever locked you out
   passwd root                 # or: set your key in /root/.ssh
   exit

4. 06.4

   Back in the console, choose Boot normally and confirm the instance comes up. Then thank the VNC console; it does not get enough credit.