`Unknown client name` on test-azure-ubuntu2204_container-arm64 nodes on ci.nodejs.org

[sxa]

Suspect this is a mismatch somewhere on the machines but these errors are in the jenkins logs and should be addressed:

Connection #1296721 from /20.84.43.2:44600 failed: null
Aug 21, 2026 6:01:29 AM INFO hudson.TcpSlaveAgentListener$ConnectionHandler run
Accepted JNLP4-connect connection #1296722 from /20.84.43.2:44610
Aug 21, 2026 6:01:30 AM INFO org.jenkinsci.remoting.protocol.impl.ConnectionHeadersFilterLayer onRecv
[JNLP4-connect connection from 20.84.43.2/20.84.43.2:44610] Refusing headers from remote: Unknown client name: test-azure-ubuntu2204_container-arm64-2
Aug 21, 2026 6:01:31 AM INFO hudson.TcpSlaveAgentListener$ConnectionHandler run
Connection #1296723 from /4.242.42.14:47498 failed: null
Aug 21, 2026 6:01:31 AM INFO hudson.TcpSlaveAgentListener$ConnectionHandler run
Accepted JNLP4-connect connection #1296724 from /4.242.42.14:47500
Aug 21, 2026 6:01:31 AM INFO org.jenkinsci.remoting.protocol.impl.ConnectionHeadersFilterLayer onRecv
[JNLP4-connect connection from 4.242.42.14/4.242.42.14:47500] Refusing headers from remote: Unknown client name: test-azure-ubuntu2204_container-arm64-1
Aug 21, 2026 6:01:32 AM INFO hudson.TcpSlaveAgentListener$ConnectionHandler run
Connection #1296725 from /4.242.42.14:52832 failed: null
Aug 21, 2026 6:01:32 AM INFO hudson.TcpSlaveAgentListener$ConnectionHandler run
Accepted JNLP4-connect connection #1296726 from /4.242.42.14:52836
Aug 21, 2026 6:01:32 AM INFO org.jenkinsci.remoting.protocol.impl.ConnectionHeadersFilterLayer onRecv
[JNLP4-connect connection from 4.242.42.14/4.242.42.14:52836] Refusing headers from remote: Unknown client name: test-azure-ubuntu2204_container-arm64-3

[richardlau]

Maybe they're supposed to be named ubuntu2404?

[richardlau]

Actually it looks like the host is running e.g. test-azure-ubuntu2404_container-arm64-2 and test-azure-ubuntu2204_container-arm64-2, however there is no Jenkins agent defined for test-azure-ubuntu2204_container-arm64-2.

nodejs@test-azure-ubuntu2404-docker-arm64-2:~$  sudo docker ps
CONTAINER ID   IMAGE                                              COMMAND                  CREATED       STATUS       PORTS     NAMES
f207c61df2f3   node-ci:test-azure-rhel8_container-arm64-2         "/bin/sh -c 'cd /hom…"   3 weeks ago   Up 3 weeks             node-ci-test-azure-rhel8_container-arm64-2
9294014a9989   node-ci:test-azure-ubuntu2404_container-arm64-4    "/bin/sh -c 'cd /hom…"   3 weeks ago   Up 3 weeks             node-ci-test-azure-ubuntu2404_container-arm64-4
b0fe71491071   node-ci:test-azure-ubuntu2404_container-arm64-2    "/bin/sh -c 'cd /hom…"   3 weeks ago   Up 3 weeks             node-ci-test-azure-ubuntu2404_container-arm64-2
2b5a126f3877   node-ci:test-azure-ubuntu2204_container-armv7l-2   "/bin/sh -c 'cd /hom…"   3 weeks ago   Up 3 weeks             node-ci-test-azure-ubuntu2204_container-armv7l-2
0789d5474ffe   node-ci:test-azure-ubuntu2204_container-arm64-2    "/bin/sh -c 'cd /hom…"   3 weeks ago   Up 3 weeks             node-ci-test-azure-ubuntu2204_container-arm64-2
0775cb734bae   node-ci:test-azure-debian13_container-armv7l-2     "/bin/sh -c 'cd /hom…"   3 weeks ago   Up 3 weeks             node-ci-test-azure-debian13_container-armv7l-2
5625a1d1392d   node-ci:test-azure-ubuntu2204_container-arm64-4    "/bin/sh -c 'cd /hom…"   3 weeks ago   Up 3 weeks             node-ci-test-azure-ubuntu2204_container-arm64-4
nodejs@test-azure-ubuntu2404-docker-arm64-2:~$

[richardlau]

There's a mismatch between what is defined in Ansible/Jenkins and what is actually running on the machines

        ubuntu2404_docker-arm64-2:
          server_jobs: 2
          containers:
            - { name: 'test-azure-debian13_container-armv7l-2', os: 'debian13_armv7l', secret: '<redacted>' }
            - { name: 'test-azure-rhel8_container-arm64-2', os: 'rhel8', secret: '<redacted> }
            - { name: 'test-azure-ubuntu2204_container-armv7l-2', os: 'ubuntu2204_armv7l', secret: '<redacted>' }
            - { name: 'test-azure-ubuntu2404_container-arm64-2', os: 'ubuntu2404', secret: '<redacted>' }
            - { name: 'test-azure-ubuntu2404_container-arm64-4', os: 'ubuntu2404', secret: '<redacted>' }

Suggests we should only be running five containers instead of the seven currently running on the machine. My guess would be that we were running Ubuntu 22.04 arm64 containters before and updated them to Ubuntu 24.04 but the Ubuntu 22.04 containers/systemd config were left behind on the system.

[richardlau]

I've stopped and cleared up the Ubuntu 22.04 arm64 systemd services on the test-azure-ubuntu2404-docker-arm64-* Docker hosts.

Generalized procedure (as root). Run systemctl status jenkins-<name_of_container>.service at any point to check the status.

systemctl stop jenkins-<name_of_container>.service
systemctl disable jenkins-<name_of_container>.service
rm /lib/systemd/system/jenkins-<name_of_container>.service
systemctl reset-failed

There were also systemd service files for

• jenkins-test-azure-debian11_container-armv7l-1.service
• jenkins-test-azure-debian11_container-armv7l-2.service

but those were already disabled (and not running). I've also removed those (looks like those containers were replaced with Debian 13 armv7l, which are running).

[richardlau]

As an aside, all three test-azure-ubuntu2404-docker-arm64-* hosts display this when logging in:

*** System restart required ***

I haven't restarted the machines as they are currently running jobs.

Update: the arm builds were quiet, so I've restarted the machines.