Get Started: Plan Your PAM Jump Server Deployment
The Jump Server is a proxy that you install on a Linux host inside your own network. Every session to a server, a database, or a Privileged Website with Web Shield goes through it. A Jump Server can show Online and still fail to serve sessions when the host, Docker, or the network is not ready.
Use this article to prepare the host and the network before you install. A few minutes of planning avoids most failed installations. When everything on the checklist is ready, continue with Install PAM Jump Server.
Planning Checklist
-
The host meets the system requirements.
-
You decided how users reach the Jump Server, and which Web Port and Shell Port to use.
-
The firewall allows users to reach the host, and allows the host to reach Privileged Access Management (PAM) and your resources.
-
If the host is an Amazon EC2 instance, the VPC and the security groups are ready, including the S3 endpoint policy.
-
You ran the readiness checks on the host.
Preparing the Host
Install the Jump Server on a dedicated Linux host. A dedicated host avoids conflicts with other services that use the same ports. The host can be a virtual machine in your data center or in a cloud provider.
| Requirement | Value |
|---|---|
| Operating system | Ubuntu Server 24.04 |
| CPU | At least 2 vCPUs |
| Memory | At least 8 GB of RAM |
| Disk | At least 80 GB free on the root file system (/). Provision more than 80 GB in total, so that the operating system and system files still leave 80 GB free. |
| Docker | Installed and running before you run the installer |
| Privileges | Root access. The installer must run as root. |
| Tools | curl available on the host |
| Clock | Within 300 seconds of the server time |
The requirements above are the minimum. As the number of concurrent sessions grows, you can upgrade the hardware of the host to keep the quality of the sessions.
Deciding How Users Reach the Jump Server
Users reach the Jump Server from their own computers. You decide whether the host has a public IP address, a private IP address, or both. A common setup is a private IP address only, with users connecting through a VPN or another access tool. Keep the path as simple as possible, because every extra layer, such as a VPN plus a ZTNA tool plus several firewalls, adds latency and makes problems harder to find.
When you add the Jump Server in PAM, you enter the IPv4 address that users reach, and two ports:
| Port | Default | Used for |
|---|---|---|
| Web Port | 443 | Browser sessions |
| Shell Port | 2222 | Direct Access |
Change the default ports only if another service on the host already uses them. The two ports are independent, so allow both in your firewall.
Preparing the Network
Inbound traffic
Allow the computers of your users to reach the host on the Web Port and on the Shell Port. You only need to allow your own users. JumpCloud never connects to the Jump Server from the internet, so you do not need to allow any JumpCloud IP address inbound.
Outbound traffic to PAM
The Jump Server starts all the connections to PAM. Allow the host to reach the PAM endpoints for your data center, which include the PAM domains, the container images, the session recording storage, and the message queue. The full list by data center is in Install PAM Jump Server.
The message queue uses port 5671, and all the other endpoints use port 443. A rule that allows only outbound port 443 is not enough. Without port 5671, the Jump Server installs and shows Online, but actions from PAM, such as disconnecting a user from an active session, do not work.
Outbound traffic to your resources
The Jump Server must reach each resource on the port of its protocol, such as 22 for SSH, 3389 for RDP, or the port of the database. Make sure the resource also accepts connections from the IP address of the Jump Server. The resource sees the IP address of the Jump Server, not the IP address of the user. If a firewall, a proxy, or a security policy between the host and PAM inspects or filters the traffic, confirm that it allows the endpoints above.
Preparing an Amazon EC2 Instance
If you install the Jump Server on an Amazon EC2 instance, also check these items in your AWS account.
-
Instance: use an Ubuntu Server 24.04 image and an instance type with at least 2 vCPUs and 8 GiB of memory. In the root volume, provision more than 80 GB, for example 100 GB.
-
Security group of the Jump Server: allow inbound TCP on the Web Port and on the Shell Port from the networks of your users. Allow outbound traffic to the PAM endpoints, including TCP 5671.
-
Security groups of your resources: allow inbound traffic from the IP address of the Jump Server on the port of each protocol. A rule that allows only your office network blocks the Jump Server, and the connection fails without a clear error.
-
IP address: the public IP address of an instance can change when you stop and start it. If users reach the Jump Server through a public IP address, use an Elastic IP address, and enter that address in PAM.
-
Outbound path in a private subnet: a gateway endpoint for Amazon S3 only carries the traffic to S3. The Jump Server also needs an outbound path to the other PAM endpoints, for example through a NAT gateway or a proxy.
Allowing the PAM bucket in the S3 endpoint policy
If your VPC has a gateway endpoint for Amazon S3 with a restrictive policy, the policy can block the Jump Server from downloading its SSL certificate. The installation still finishes, but the certificate is not installed, and users see a browser warning when they open the address of the Jump Server.
To avoid it, add the PAM bucket to the policy of the endpoint before you install. In VPC, select Endpoints, open the endpoint with the service name com.amazonaws.
arn:aws:s3:::vaultone-prd-data
arn:aws:s3:::vaultone-prd-data/*
If your policy does not restrict buckets, you do not need to change it. If the certificate is already blocked, see Troubleshoot: Error 403 Amazon S3 Forbidden.
Running the Readiness Checks
Run these commands on the host, and compare the results with the requirements. The installer runs similar checks, and it stops and names the check that failed.
- Check the operating system. The result must show Ubuntu 24.04.
cat /etc/os-release
- Check the CPUs, the memory, and the free disk space of the root file system.
nproc
free -h
df -h /
- Check that the clock is synchronized.
timedatectl
- Check that Docker and curl are installed, and that Docker is running.
docker --version
systemctl status docker
curl --version
- Check the access to the container images. Any HTTP response means that the host reaches the endpoint.
curl -I https://public.ecr.aws
The result must say that the connection succeeded. A timeout means that a firewall blocks the port.
Next Step
When the checklist is complete, register the Jump Server in PAM and run the installer. In the Admin Portal, go to Access > PAM > Jump Servers. For the steps, see Install PAM Jump Server.
Was this information helpful?