Start your 3-day free trial
Sign up to experience all premium features at no cost.
*Available only to new users. Each user is limited to one trial.


A NoMachine connection timeout means a connection did not complete within the expected time; it does not identify the failed component. Confirm the saved host address, the remote service and the permitted access path before changing your network. Recovery means reaching the intended computer, signing in with authorization and using its desktop.
Key Takeaways:
- Separate an IP-based connection from NoMachine Network before troubleshooting.
- Verify the remote listener and the exact external port with the host owner.
- A refusal, timeout and sign-in failure require different next actions.
- Compare client network paths only for an already approved public target.
NoMachine's getting-started guide distinguishes connections using a computer's IP address from NoMachine Network. The latter is a separate service intended for reaching machines without a public address or behind NAT. This article follows an existing IP-based connection; do not apply its public-address comparison to a machine ID or a managed gateway. [1]
| Observation | Next action | Boundary |
|---|---|---|
| Saved host is a private address | Restore the approved LAN or enterprise access path | A consumer VPN does not supply that private route |
| Hostname or address differs from the owner's record | Verify the connection entry with the owner | Do not guess a new destination |
| Connection is refused | Check the selected service, listener and filtering | A refusal is not a password verdict |
| Connection times out before sign-in | Check host availability and the existing route | A timeout does not prove ISP blocking |
| Sign-in or desktop selection appears | Check credentials, consent and desktop availability | A working network does not grant permission |
Keep the investigation small. Name the computer you want to control and the computer you are using, then write down which one displays the error. A screenshot containing credentials, a workplace hostname or a public address should stay in the approved support channel. For a wider device investigation, use the broader checklist for device and app faults.
Open the saved NoMachine connection without editing it. Record the host, protocol, port, client version, remote version if known, and whether access is direct by IP, through an organization-provided path, or through NoMachine Network. Ask the owner to confirm the same tuple. A familiar connection name can survive an address change and point to the wrong host.
For a LAN connection, obtain the remote address from the host's NoMachine status information. For an Internet connection, the official guide uses the public-facing address and its associated external port. The default NX service port is 4000, but an existing external mapping can use another port; copy the verified connection information rather than assuming every saved connection should use 4000. [1]
Treat these as different destinations even when they belong to the same computer. A private address that worked while you were in the office does not become a public endpoint after you leave. Ask the administrator how that computer is meant to be reached from your location. Do not replace its address with the IP shown by a random website on your client.
Before another attempt, confirm that the host owner expects you, that the computer is powered on, and that the intended account is permitted to connect. If you cannot establish ownership or permission, stop. The aim is to repair a known connection, not discover machines on an unfamiliar network.
For a workplace host, use the workplace VPN permission checks to clarify the designated access method with IT before testing another path.
Have the owner check the remote computer locally or through an existing approved management channel. Verify that the NoMachine server is running and that the selected service uses the port recorded in the connection. The official refusal article specifically directs readers to check address, port, firewall configuration and server status. [2]
Its screenshots and menu descriptions refer to older versions, including v5 and v6. Use the equivalent service and status controls in the installed version instead of looking for an identical tab. The current getting-started guide presents server status and NX configuration separately. Record what your installation actually shows, including its edition, rather than transplanting a menu sequence from another release. [1][2]
A useful owner report contains three observations: the host is awake, the server is running, and the selected protocol and port match the intended connection. If the owner has a permitted local client available, a session to that same host can help check the host side. Success on the LAN narrows the investigation, but it does not validate the external address or mapping.
Do not restart a server carrying other users' sessions just to obtain a cleaner result. Ask the owner or administrator to schedule a controlled action if the service is stopped or the configuration is inconsistent. Leave router mapping and firewall changes with whoever owns those controls; this workflow does not require disabling protection or creating a new public exposure.
Start one attempt against the verified entry and record its time, full error and last visible stage. Did the failure occur before any account prompt, after you submitted credentials, or after the desktop appeared? Those observations are more useful than a general report that NoMachine remote access failed.
For NoMachine connection refused, first revisit the listening service and the port. The vendor's article also calls for checking both sides' firewall configuration and, if the problem persists, reviewing logs from both endpoints. It does not say that every refusal proves a firewall rule is responsible. [2]
For a timeout before sign-in, request a contemporaneous host check and confirm that the existing access path is still intended to work. A host that went to sleep between attempts creates a different comparison from two attempts made while it stayed awake. Keep your observation separate from your explanation: write “no account prompt appeared” before suggesting a route problem.
If sign-in appears, proceed only with the correct remote system account and any required consent. The official IP-based workflow uses credentials for an account on the destination computer. A NoMachine website account is not automatically that system account. [1] A rejected credential or unavailable desktop should be investigated at that stage, not treated as proof that you need a different Internet exit.
Avoid repeated retries while changing addresses, credentials and networks together. Preserve one connection entry, make one justified correction at a time, and record the next outcome. If the approved network policy forbids the connection, stop and request the designated access method.
If the access computer uses macOS and its proxy configuration recently changed, review the Mac proxy settings and their scope. Preserve required settings and obtain the owner's approval before correcting them; browser access alone does not validate the NoMachine connection.
This branch is appropriate only when the owner has already approved public access, the public destination and external port are verified, and the host is awake with the service running. It does not create public access. Do not enable automatic port mapping merely to make this test possible; an administrator must decide how the destination is exposed and protected.
First attempt the unchanged connection from the client without a consumer VPN. Record whether it reaches sign-in and whether a desktop session succeeds. Then, if local rules and organizational policy permit the comparison, change only the client's Internet path. Leave the remote network, host configuration, credentials and saved connection tuple unchanged.
Use the VPN checks before a NoMachine session to confirm the intended client path, then judge the NoMachine result by the authorized desktop task.
For that second attempt, you can use AethoVPN on a supported client computer with Pro or Premium. On Windows, install the official Windows package; on macOS, obtain the configuration through the official setup guide and connect using a VPN client. Linux installation is available as a .deb for Debian/Ubuntu x64. Select a nearby currently available location, enable global mode so application traffic follows the comparison path, and reconnect the same NoMachine entry.
Keep the two records together: client path, attempt time, last stage, error, remote owner confirmation, and desktop result. If only the second path reaches a usable authorized desktop, you have evidence of a difference between those paths for that target at that time. You have not established the exact failing router, a general ISP policy, or a permanent fix. If neither path works, return to the host and route evidence instead of cycling through locations.
AethoVPN cannot supply the remote host's public IP, create its port mapping, grant access to an enterprise private network, or guarantee LAN reachability. Global mode can change access to existing private resources, so stop this branch for a private destination and restore the organization's designated method. Also stop if a source-address restriction rejects the new path or an administrator disallows it; do not weaken the restriction to continue.
If this permitted public-path comparison fits your case, start the 3-day Pro trial. The trial begins at registration and is available once per user. Registration uses an email verification code; check that the client platform and current location list suit the experiment before paying.
A port response or a login dialog is an intermediate result. Confirm the identity of the remote computer through an owner-approved detail, sign in with the authorized account, and complete a harmless task such as opening the intended work folder and checking that input updates the screen. Do not access another person's files to demonstrate recovery.
Record whether the session stays usable for the task that originally failed. If the desktop appears but input is rejected, a display is unavailable, or permission is missing, describe that specific stage to the owner. Do not label the whole connection repaired because its error message changed.
End the session, restore the original client path if the comparison was temporary, and retain the two attempt records in the approved channel. Avoid leaving the comparison VPN enabled while testing a different private destination: that would change the question being investigated. If a new configuration is adopted, document its owner and approved purpose separately.
When handing the case to support, include both versions and editions, the deployment type, redacted destination and port details, attempt times, service observations and the last successful stage. Logs from both endpoints may be needed, as the vendor's refusal guidance explains. [2] Remove credentials and unrelated personal information before sharing; send the record to the actual maintainer, not a public comment thread.
Check whether it contains a private address. That entry may require the office LAN or the organization's approved remote-access path. Changing your client Internet exit does not recreate that route; ask the administrator which endpoint you should use from your current location.
Only if the owner confirms it. NX uses 4000 by default, but an Internet entry can need a different external port. Compare the saved entry with the verified host and mapping information; do not overwrite a deliberate configuration because a tutorial shows the default. [1]
It does not establish that. Check the destination, listener and filtering first. A refusal before an account prompt and a credential rejection after submitting an account occur at different stages and should be reported separately. [2]
A local session checks a local path. It does not validate the public-facing destination or existing mapping used from outside. Give the owner both attempt records and let them check the approved external path; do not create a new mapping without their authorization.
This guide does not use it that way. A private destination needs the organization's designated routing and access controls. The comparison described here changes a client's path to an already permitted public endpoint; it supplies neither enterprise authorization nor a private-network route.
Reach the intended host, authenticate with permission, and perform the agreed desktop task. A changed IP address, open vendor website, successful port probe or visible sign-in prompt alone is insufficient. Record the path and conditions so the result can be interpreted correctly.
Stop for an unknown destination, missing permission, unavailable host, inconsistent service configuration or a policy restriction. Also escalate if both permitted paths fail with the same verified tuple. Provide the last successful stage and redacted endpoint observations rather than claiming that a VPN must fix the timeout.
Sources:
Sources checked 7 October 2026.
Related Articles:
Sign up to experience all premium features at no cost.
*Available only to new users. Each user is limited to one trial.





