VPS Guides
Tested on Ubuntu 24.04 · Debian 12 · AlmaLinux 9
Documentation · Version 3.0
VPS guides — from blank server to live site.
Ten hands-on walkthroughs for the work every new server needs: first login, SSH keys, firewall hardening, a LEMP stack, WordPress and WooCommerce, free TLS, snapshots, resizing, migration and API automation. Each guide states its time, difficulty and which of our sixteen OS templates it applies to. Commands are copy-ready; placeholders are marked in italics so you never paste a template literally.
Reviewed 29 October 2026 · Applies to all WenHost VPS plans
- No paid tools required
- Panel steps named exactly
- Policy limits flagged inline
Before you start
Five things make every guide below work the first time. Two of them are policy limits customers most often trip over, so they are stated here rather than buried.
| You need | Where it comes from |
|---|---|
| Public IPv4 of your VPS | Delivery email and client area, assigned during provisioning. |
| Root or Administrator password | Delivery email. Rotate it on first login — guide 1. |
| Client-area login | Same email; console, snapshots and firewall live there. |
| A domain with DNS you control | Yours. Point an A record at the IPv4 before guide 6. |
| Terminal with SSH | macOS or Linux built in; on Windows use PowerShell or PuTTY. |
Two limits to know now. Outbound ports 25 and 587 are blocked by default on every instance — mail sending needs prior approval under clause 6 of the Acceptable Use Policy. And plans are unmanaged: we keep the platform healthy and diagnose to the OS boundary, but your software is yours to administer, per clause 4 of the Terms of Service.
Connect for the first time
- 4 minutes
- Beginner
- Linux templates
- Panel + terminal
Your instance is already running — delivery completed when it answered our probe and the credentials reached your mailbox, the definition in clause 3 of the Delivery Policy. This guide gets you in and changes the one secret you should never leave as shipped.
-
Open the terminal and connect
Replace the address with your own IPv4 from the delivery email.
your laptopssh root@203.0.113.42 # first connection asks you to trust the host key — type yes # then paste the root password from the delivery email -
Confirm you are on the machine you paid for
root@vpshostnamectl # check the CPU, memory and disk match your plan df -h / free -hIf the specification does not match the plan you ordered, say so within 72 hours of activation — clause 12 of the Delivery Policy gives you a correction, a pro-rata credit or a full refund, and it does not consume your money-back guarantee.
-
Rotate the root password immediately
root@vpspasswd # 20+ random characters; store it in your password manager -
Create a daily-use user with sudo
Working as root for routine tasks is how servers get broken.
root@vpsadduser deploy usermod -aG sudo deploy # on AlmaLinux, Rocky, CentOS or Fedora use: usermod -aG wheel deploy exit -
Log in as the new user
your laptopssh deploy@203.0.113.42 sudo whoami # prints root — sudo works, you are ready for guide 2
Cannot connect at all? Do not wait. Open the VNC console in the client area — it works even when SSH is broken — and check the firewall. If the instance never answered a probe, that is a delivery failure and clause 12 of the Delivery Policy applies from the moment you report it.
Set up SSH keys and disable password login
- 6 minutes
- Beginner
- Linux templates
- Biggest security win
Internet-facing servers with password SSH are probed within minutes — that is exactly the brute-force traffic our abuse desk handles. A key pair removes the whole attack class, and it is faster to type nothing than to type a password.
-
Generate a key pair on your own computer
Never generate private keys on the server and download them.
your laptopssh-keygen -t ed25519 -C "[email protected]" # press Enter for the default path, then set a passphrase -
Copy the public key to the server
your laptopssh-copy-id deploy@203.0.113.42 ssh deploy@203.0.113.42 # should log you in without a password prompt -
Keep the VNC console open before the next step
This is the one step where people lock themselves out. Leave a console session authenticated in another browser tab so you can undo a mistake.
-
Disable password authentication
deploy@vpsprintf 'PasswordAuthentication no\nPubkeyAuthentication yes\nPermitRootLogin no\n' | sudo tee /etc/ssh/sshd_config.d/10-hardening.conf sudo sshd -t # no output means the config is valid — only then restart sudo systemctl restart ssh # on RHEL-family systems the service is named sshd -
Verify from a second terminal
your laptopssh -o PubkeyAuthentication=no deploy@203.0.113.42 # must be refused — password login is now closedDo the same for the
rootuser:PermitRootLogin nomeans administration happens as your user withsudo, and every privileged action is attributable.
Lost your key? Use the VNC console in the client area to add a new public key to ~/.ssh/authorized_keys. We can reset a console password after verifying your identity against the payment record — the same check described in clause 2 of the Terms of Service, which protects you from account takeover.
Harden the firewall and block brute force
- 5 minutes
- Beginner
- Linux templates
- Required by AUP clause 14
Clause 14 of the Acceptable Use Policy makes host firewalling and patching your obligation, because an exposed Redis or memcached port becomes a DDoS amplifier that hurts strangers. This guide closes the door and adds a bouncer.
-
Set default-deny inbound, allow outbound
deploy@vpssudo ufw default deny incoming sudo ufw default allow outgoing sudo ufw allow OpenSSH sudo ufw allow 80/tcp sudo ufw allow 443/tcp sudo ufw enable sudo ufw status verboseOn AlmaLinux, Rocky, CentOS and Fedora use firewalld instead:
deploy@vpssudo firewall-cmd --permanent --add-service=ssh sudo firewall-cmd --permanent --add-service=http --add-service=https sudo firewall-cmd --reload -
Open only what a service actually needs
Add a rule when you install something that listens — database, game server, panel — and bind management services to
localhostor your private VLAN rather than the public interface.deploy@vpssudo ss -tulnp # every LISTEN line must be one you chose deliberately -
Install fail2ban
deploy@vpssudo apt install fail2ban -y sudo cp /etc/fail2ban/jail.conf /etc/fail2ban/jail.local sudo systemctl enable --now fail2ban sudo fail2ban-client status sshdOn RHEL-family systems install from EPEL:
sudo dnf install epel-release && sudo dnf install fail2ban. -
Turn on automatic security updates
deploy@vpssudo apt install unattended-upgrades -y sudo dpkg-reconfigure -plow unattended-upgrades # RHEL family: sudo dnf install dnf-automatic && sudo systemctl enable --now dnf-automatic-install.timer -
Use the panel firewall as a second layer
The client area has a stateful edge firewall in front of the instance. Host rules and edge rules both apply, so an accidental
ufw disabledoes not expose everything at once.
Install a LEMP stack on Ubuntu 24.04
- 12 minutes
- Intermediate
- Ubuntu · Debian
- Nginx + MySQL + PHP
Linux, Nginx, MySQL and PHP — the stack behind most WordPress and WooCommerce deployments. Commands are current for Ubuntu 24.04 with PHP 8.3.
-
Update first, always
deploy@vpssudo apt update && sudo apt upgrade -y sudo reboot -
Install Nginx, MySQL and PHP
deploy@vpssudo apt install nginx mysql-server php-fpm php-mysql \ php-xml php-mbstring php-curl php-zip php-gd php-intl -y php -v systemctl status php8.3-fpm --no-pager -
Secure MySQL and create a database
deploy@vpssudo mysql_secure_installation # set a root password, remove anonymous users, disallow remote root sudo mysql -e "CREATE DATABASE wp_demo CHARACTER SET utf8mb4 COLLATE utf8mb4_unicode_ci;" sudo mysql -e "CREATE USER 'wpuser'@'localhost' IDENTIFIED BY 'strong-password';" sudo mysql -e "GRANT ALL PRIVILEGES ON wp_demo.* TO 'wpuser'@'localhost'; FLUSH PRIVILEGES;"Keep the database on
localhost. If you must reach it from another instance, use a private VLAN — included on VPS Pro and above — rather than opening 3306 to the internet. -
Add an Nginx server block
/etc/nginx/sites-available/demoserver { listen 80; listen [::]:80; server_name shop.example.com; root /var/www/demo/public; index index.php index.html; location / { try_files $uri $uri/ /index.php?$args; } location ~ \.php$ { include snippets/fastcgi-php.conf; fastcgi_pass unix:/run/php/php8.3-fpm.sock; } location ~ /\.ht { deny all; } }deploy@vpssudo mkdir -p /var/www/demo/public sudo chown -R www-data:www-data /var/www/demo sudo ln -s /etc/nginx/sites-available/demo /etc/nginx/sites-enabled/ sudo nginx -t && sudo systemctl reload nginx -
Point your domain at the server
Create an
Arecord forshop.example.comat your DNS provider with your VPS IPv4 as the value, and lower the TTL to 300 seconds before you later switch hosts. Set the reverse DNS (PTR) for your IP in the client area so mail and monitoring see a consistent name — clause 8 of the Delivery Policy explains how PTR is assigned. -
Confirm it answers
your laptopcurl -I http://shop.example.com # expect HTTP 1.1 200 OK from nginx
Deploy WordPress and WooCommerce with WP-CLI
- 15 minutes
- Intermediate
- All Linux templates
- No manual uploads
WP-CLI installs, configures and updates WordPress from the terminal — repeatable, scriptable, and it avoids the FTP credentials that many hosts force on you.
-
Install WP-CLI
deploy@vpscurl -sO https://raw.githubusercontent.com/wp-cli/builds/gh-pages/phar/wp-cli.phar chmod +x wp-cli.phar sudo mv wp-cli.phar /usr/local/bin/wp wp --info -
Download WordPress into the web root
deploy@vpscd /var/www/demo/public sudo -u www-data wp core download --version=latest -
Create the configuration and install
deploy@vpssudo -u www-data wp config create \ --dbname=wp_demo --dbuser=wpuser --dbpass='strong-password' --dbhost=localhost sudo -u www-data wp core install \ --url="https://shop.example.com" --title="Demo Store" \ --admin_user=admin --admin_password='strong-password' \ --admin_email=[email protected] --skip-email -
Add WooCommerce and a store theme
deploy@vpssudo -u www-data wp plugin install woocommerce --activate sudo -u www-data wp theme install storefront --activate sudo -u www-data wp rewrite structure '/%postname%/' --hard -
Set sane permissions and finish in the browser
deploy@vpscd /var/www/demo/public sudo find . -type d -exec chmod 755 {} \; sudo find . -type f -exec chmod 644 {} \; sudo chown -R www-data:www-data wp-contentComplete the WooCommerce setup wizard, then add a page cache and an object cache. On a 2 GB plan start with PHP OPcache plus a full-page cache; you do not need Redis on day one.
-
Keep it updated safely
deploy@vpssudo -u www-data wp core update-db sudo -u www-data wp plugin update --all crontab -l | sudo -u www-data tee -a /dev/null # schedule: 0 4 * * * cd /var/www/demo/public && wp plugin update --all --quietTake a snapshot in the client area before every update batch. Snapshots are free and kept 7 days — guide 7 shows the restore path.
Payment gateways on your store: WenHost charges you in USD by PayPal, and your own customers pay your store through whichever gateway you integrate. We do not intermediate your store's payments, and we never see your gateway API keys. Keep them out of wp-config.php in version control.
Issue a free TLS certificate
- 4 minutes
- Beginner
- All Linux templates
- Auto-renews
Let's Encrypt certificates are free and valid 90 days; certbot renews them automatically. Do this only after the A record resolves to your server, or validation fails.
-
Confirm DNS has propagated
deploy@vpsdig +short shop.example.com # must return your VPS IPv4 -
Install certbot and issue
deploy@vpssudo apt install certbot python3-certbot-nginx -y sudo certbot --nginx -d shop.example.com -d www.shop.example.com # choose the redirect option so HTTP always goes to HTTPS -
Verify the renewal timer
deploy@vpssudo certbot renew --dry-run systemctl list-timers | grep certbot -
Update the site address
deploy@vpscd /var/www/demo/public sudo -u www-data wp search-replace 'http://shop.example.com' 'https://shop.example.com' -
Check the chain and grade
Load the site over HTTPS, confirm the padlock, and check that port 443 is open in both your host firewall and the panel firewall. If a browser reports a mixed-content warning, the certificate is fine — a hardcoded
http://asset needs replacing.
Snapshots, restore and reinstall
- 3 minutes
- Beginner
- All templates
- Panel only
Daily snapshots are included on every plan and kept 7 days. They are our hardware safety net and your rollback button — but clause 11 of the Terms of Service is explicit that they do not replace your own off-platform backup.
-
Take a manual snapshot before risky work
Client area → Instances → select your server → Snapshots → Create. Name it after the change you are about to make, for example
before-nginx-upgrade. -
Restore in one click
Snapshots → choose the point → Restore. The instance is briefly offline during restore. Everything on the disk returns to that moment, including mistakes made since — which is why naming matters.
-
Enable off-site copies if you need them
By default snapshots live in the same region as the server. Turn on off-site replication for disaster recovery against a regional event.
-
Reinstall to a different OS — carefully
Reinstalling erases the disk. The panel warns you and offers a snapshot first; take it. A reinstall is a fresh delivery of the same instance, so the activation timestamp for that instance updates while your billing cycle does not — clause 11 of the Delivery Policy.
Client area → Reinstall → pick a template from the sixteen available → confirm. One to three minutes typical, one hour committed.
-
Rescue mode for a broken configuration
Boot a live environment from the panel to mount your disk, fix a bad
sshd_configor a broken fstab, then reboot normally. No data is lost. -
Build your own backup habit
deploy@vpssudo mysqldump --single-transaction wp_demo | gzip > /backup/wp-$(date +%F).sql.gz tar czf /backup/site-$(date +%F).tar.gz /var/www/demo/public # then copy /backup somewhere that is not this server
Resize your plan without losing data
- 5 minutes
- Beginner
- All templates
- Prorated billing
CPU, RAM and disk grow from the panel in under five minutes with one reboot, and billing is prorated for the remainder of the period — clause 6 of the Terms of Service.
-
Snapshot first
Free, 30 seconds, and the reason a resize is a non-event rather than a gamble.
-
Choose the new size
Client area → Resize → select the plan → review the prorated difference shown before you confirm. Growing never requires a rebuild.
-
Reboot and extend the filesystem
deploy@vpssudo reboot # after it returns: sudo growpart /dev/vda 1 sudo resize2fs /dev/vda1 # XFS instead of ext4? use: sudo xfs_growfs / df -h / -
Shrinking is different
Disk cannot be reduced in place. A downgrade needs a reinstall or a fresh instance plus a migration, and the panel tells you before you commit. Credits for a downgrade go to your account balance, not to your card.
-
Know when to move to a dedicated host instead
If you are consistently at maximum CPU on eight cores, or your I/O is degrading the shared array, more of the same plan is the wrong answer. Clause 8 of the Acceptable Use Policy covers fair use — we will contact you with numbers and options rather than throttling silently.
Migrate from another host for free
- 24 hours
- Intermediate
- All templates
- One server included
One migration per new VPS subscription is free. The clock starts when you supply access and approve the plan, not when you pay — clause 10 of the Delivery Policy.
-
Open the migration request
Use the contact form with the topic Migration from another host. Include the source IP or panel URL, the accounts and databases to move, and a maintenance window you accept. Never send credentials by ordinary email to a personal address.
-
We copy and test — your live site untouched
Within 24 hours the data is on your new VPS, booted on a temporary hostname so you can click through everything: logins, checkout, cron, mail routing, API endpoints.
-
Fix the URLs before cutover
deploy@vpssudo -u www-data wp search-replace 'http://old-host.example' 'https://shop.example.com' sudo -u www-data wp cache flush -
Lower the DNS TTL, then point traffic
At your registrar set the TTL to 300 seconds an hour ahead, then change the A record. Propagation typically completes within 24 hours and depends on your DNS provider, not on us.
-
Keep the source alive for 7 days
We leave the old server untouched for a week so rollback is trivial. Do not cancel it with your previous host until you have confirmed mail, cron jobs and outbound integrations all work from the new IP.
-
Complex moves get a schedule, not a promise
Many accounts, large mailboxes, databases over 100 GB, or platforms without standard APIs are quoted a specific plan instead of the 24-hour target. The commitment for a standard single-server move is 3 business days.
Automate with the REST API and Terraform
- 10 minutes
- Advanced
- All templates
- IaC ready
Everything the panel does is available over the API — that is how provisioning reaches 45 seconds. Create a scoped token in the client area first, and treat it like a password.
-
Authenticate and list instances
your laptopexport WENHOST_TOKEN=wh_live_xxxxxxxxxxxx curl -s -H "Authorization: Bearer $WENHOST_TOKEN" \ https://api.wenhost.website/v1/instances | jq '.[].name' -
Create a server from code
your laptopcurl -s -X POST https://api.wenhost.website/v1/instances \ -H "Authorization: Bearer $WENHOST_TOKEN" \ -H "Content-Type: application/json" \ -d '{ "plan": "vps-pro", "region": "singapore", "template": "ubuntu-24.04-x64", "ssh_keys": ["ssh-ed25519 AAAC... [email protected]"], "label": "staging-01" }'Pass an SSH key at creation and skip password login entirely — the API accepts keys so a freshly built server is hardened from its first second.
-
Terraform provider
main.tfterraform { required_providers { wenhost = { source = "wenhost/wenhost" version = "~> 1.0" } } } resource "wenhost_instance" "web" { name = "web-01" plan = "vps-starter" region = "frankfurt" template = "debian-12-x64" snapshots = true } -
Rotate and scope tokens
Create one token per integration with the narrowest permission that works, and revoke immediately when a job finishes. API calls are logged with source IP under clause 3 of the Privacy Policy, so a leaked token is traceable and closable.
-
Automate the boring safety
your laptopcurl -s -X POST https://api.wenhost.website/v1/instances/id/snapshots \ -H "Authorization: Bearer $WENHOST_TOKEN" \ -d '{"name":"nightly"}'A nightly snapshot call before a deploy pipeline runs costs one line and saves an afternoon.
Windows Server: first RDP session
- 6 minutes
- Beginner
- Server 2019 · 2022 · 2025
- Licence add-on
Windows templates boot with a licensed Windows add-on billed monthly. That licence is issued by the vendor and is non-refundable once issued, per the Refund Policy — so confirm the plan fits before you reinstall repeatedly.
-
Set the Administrator password
Client area → Instances → your server → Reset password. The change applies on the next boot; reboot from the panel.
-
Open RDP in the panel firewall
Allow TCP
3389and restrict the source to your own IP. Exposing RDP to the whole internet is the single most reported abuse vector on our abuse queue. -
Connect
Windows workstationmstsc /v:203.0.113.42 # username: Administrator · password: the one you just set -
Harden immediately
Run Windows Update to completion, create a named administrator account and rename the built-in one, enable the Windows Firewall profile, and install an endpoint protection agent. Then take a snapshot as your clean baseline.
-
Change the RDP port only as a noise reducer
Moving off 3389 cuts log-volume noise but is not a security control. Key or restricted-source access is the control.
When something does not work
The eight situations support is asked about most, with the check that settles each one.
| Symptom | Check this first | Then |
|---|---|---|
| SSH times out | Is the instance Running in the panel? Is port 22 open in both firewalls? | Boot the VNC console and inspect sshd status. Platform-side fault inside 72 h is refundable under the Refund Policy. |
| Permission denied on login | Right key offered? ssh -vvv shows which key is tried. | Check ~/.ssh is 700 and authorized_keys 600, and that PermitRootLogin no was not applied before your user existed. |
| Site loads by IP, not by domain | dig +short yourdomain — does it return your IPv4? | Fix the A record or wait out the TTL. Nginx server_name must match the host header. |
| Certbot fails validation | Port 80 reachable from outside? Redirect to HTTPS breaking the HTTP-01 challenge? | Open 80, or use the DNS challenge: certbot certonly --dns-manual -d yourdomain. |
| Mail will not send | Outbound 25/587 blocked by default — this is expected. | Request approval per AUP clause 6 with your volume, list source, SPF, DKIM, DMARC and matching PTR. |
| Server slow or unresponsive | Panel metrics: CPU steal, memory, disk I/O, connection count. | Sustained saturation means the plan is too small — resize (guide 8) or ask about a dedicated host. |
| Under a DDoS attack | Status page for your region, then panel network graph. | Layer 3/4 filtering is already on. If your IP is temporarily null-routed to protect neighbours, we tell you the start and end times — Abuse Policy clause 9. |
| Renewal charge failed | Payment method valid? PayPal authorisation live? | Update payment in the client area. Grace runs to day 3, suspension day 4, release day 11 — Terms clause 9. |
What to include when you ask us. Instance IP or hostname, the exact time it started with timezone, what you were running at the time, and any output you already saw. A screenshot from the VNC console tells us more than a paragraph of description. This is the same list our Abuse Policy asks reporters for — evidence makes everything faster in both directions.
Hardening checklist before you go live
Ten minutes of work that prevents most tickets we ever receive.
✓ Do all of these
- SSH keys, password auth off.Guide 2 — the biggest single win.
- Default-deny inbound firewall.Open only the ports you serve.
- Unattended security updates on.Guide 3, step 4.
- Non-root user with sudo.Every privileged action attributable.
- Snapshot before every change.Free, 30 seconds, 7-day retention.
- Off-platform backup on a schedule.Terms clause 11 — snapshots are not your disaster recovery.
- Bind admin services to localhost or VLAN.Never 0.0.0.0 with no auth.
✕ Never do these
- Leave Redis, memcached or Elasticsearch open.Each has been used as a DDoS amplifier — AUP clause 14.
- Run an open mail relay.Your server sending other people's spam ends the account.
- Expose RDP or SSH to the whole internet with passwords.The two most brute-forced services there are.
- Reuse the shipped root password.It travelled by email; rotate it.
- Store API keys in a public repository.Rotate on sight if it happens.
- Scan networks you do not own.Prohibited by AUP clause 7, even for research.
Guide questions, answered
Short answers, each pointing at the policy that governs it.
Do you set any of this up for me?
Plans are unmanaged by default, so the software layer is yours — Terms clause 4. We keep the platform healthy, give you the console and API, and diagnose to the OS boundary with log evidence. Optional Managed Care covers patching, monitoring and hardening for a flat monthly fee.
Which Ubuntu version should a beginner pick?
Ubuntu 24.04 LTS or Debian 12. Both are long-supported, have current packages for PHP 8.3 and Nginx, and every guide on this page is tested against them. CentOS 7 is end-of-life upstream — choose AlmaLinux 9 instead.
Why is my mail not going out?
Outbound ports 25 and 587 are blocked by default on every instance. That is deliberate: it is the main defence for the shared address space every customer depends on. Ask for approval with your volume, list source and correct SPF, DKIM, DMARC and PTR — AUP clause 6.
Can I move my server to another region?
Yes. A region change is handled as a migration: typically 15–60 minutes depending on data volume, with a 24-hour commitment, and old copies destroyed on the retention schedule in Privacy clause 12 — Delivery clause 9.
Will a resize lose my data?
No. Growing CPU, RAM or disk takes under five minutes with one reboot and leaves the disk untouched. Only shrinking disk requires a rebuild, and the panel warns you first — guide 8 and Delivery clause 11.
Something broke right after delivery. Am I covered?
Yes. A server that was not delivered on time, was unreachable at handover, or does not match the specification is refundable within 72 hours of activation on every order, and that refund does not consume your one-time money-back guarantee — Delivery clause 12, Refund Policy.
How do I claim an SLA credit?
Open a ticket within 14 days of the incident. Credits are 5% or 10% of the monthly fee by severity, capped at one month, applied to your next invoice — and they are not a cash refund, so your guarantee stays unused — Terms clause 13.
Stuck between two steps?
Send us the guide number, the step, and the exact output you got. Support engineers have root-level access and answer with the log line that shows where the fault sits — even on unmanaged plans, that diagnosis is free. Median first reply is under five minutes.
Reviewed 29 October 2026. Commands are tested on Ubuntu 24.04, Debian 12 and AlmaLinux 9 and may differ on other templates or future releases. Package versions, panel labels and policy windows match the Terms of Service, Refund Policy, Delivery Policy, Acceptable Use Policy and Abuse Policy as published on this site.
