Running an unsupported version of v15.X.X was rejected. Connections will be allowed after upgrading the agent to v16.0.0 or newer

[MisterDeeds]

I recently upgraded my Teleport cluster from v15 to v16 and then to v17. After the upgrade, all Teleport clients correctly report version 17. However, the Web UI still displays the following error message:

Connection from Node XXX.SrvTeleP01 at 192.168.1.95:54048, running an unsupported version of v15.4.27 was rejected. Connections will be allowed after upgrading the agent to v16.0.0 or newer

Despite the fact that all nodes and agents are running v17, the UI still seems to reference an older version.

Steps to reproduce:
Upgrade Teleport from v15 to v16.
Upgrade from v16 to v17.
Verify that all nodes and clients report v17.
Check the Web UI for any errors.

Expected behavior:
The Web UI should recognize that all nodes are running v17 and should not report them as outdated.

Actual behavior:
The Web UI incorrectly reports that nodes are running v15.4.27 and rejects their connections.

Environment:
Teleport version: v17.2.3
OS: Debian 12
Deployment method: Plain via script from gui

Is this a known issue? Are there any additional steps required to complete the upgrade process?

Thanks!

[zmb3]

You have a node on v15 trying to connect.

It won't show up in the inventory because the connection attempts are rejected.

[MisterDeeds]

You have a node on v15 trying to connect.

It won't show up in the inventory because the connection attempts are rejected.

hi and thanks for the answer. The host mentioned with the ip 192.168.1.95 is also on version 17:

[zmb3]

This is definitely not a UI issue - that message is generated by the backend, so if you're still seeing that alert the only explanations I can think of are:

1. You just recently updated and the alert will clear itself automatically soon (this sounds unlikely based on what you've shared so far).
2. You are still running v15.4.27 on this node.

Item 2 seems most likely to me.

Your dpkg output shows that version 17.2.3 of the package is installed on the system, but it doesn't prove that v17.2.3 is actually what's running on the system. I might try a systemctl restart teleport just to see if maybe the currently running process is still the old v15 process.

[MisterDeeds]

In the meantime I have rebooted both Teleport servers as well as the target system (SrvMail01).

However, the message still appears after logging in.

I have tested it on a second teleport cluster, there is also a host whose agent is on v17 but the message appears in the GUI.

[MisterDeeds]

So the message can't be killed... I have now completely removed the teleport service on the host with the ip 192.168.1.95 and also removed all entries from the teleport server (tctl rm node/....) and the message still appears.

[zmb3]

It will disappear on its own if you truly have no v15 nodes that are trying to connect.

Again, tctl rm node is not going to help you here because the node that is being rejected will never show up in tctl nodes ls to begin with - it isn't successfully connecting.

As with any cluster alert, you can ack it to prevent it from showing up: https://goteleport.com/docs/faq/#can-i-suppress-warnings-about-available-upgrades

[MisterDeeds]

Ah cool, I have clear them using:

tctl alerts ack rejected-unsupported-connection --clear

and now it no longer appears again. Many thanks for the help!