Skip to main content

Why does a computer appear offline?

A computer may appear offline in Medulla even though it is still present in the inventory.

This status means that Medulla is no longer correctly receiving the computer’s agent presence.

There are several possible causes:

  • the computer is actually turned off or inaccessible;
  • the MedullaAgent service is no longer functioning;
  • the agent is reconfiguring itself or restarting in a loop;
  • communication with Medulla has been interrupted;
  • the computer is experiencing a registration issue;
  • there is an inconsistency in the computer’s network information.

Note:
It is recommended that you first check the client computer and its agent logs before making any changes to the Medulla server.


1. Verify that the computer is actually available

Start by verifying that the computer is powered on and accessible on the network.

From a device that can reach the computer:

ping MACHINE_IP

If the computer does not respond, check the following in particular:

  • that it is powered on;
  • that it is connected to the network;
  • that it still has a valid IP address;
  • that no network filtering is preventing it from communicating with Medulla.

Note:
A lack of response to a ping does not always mean the computer is turned off, as some environments may block ICMP. This test is only an initial check.


2. Check the Medulla agent logs

On a Windows computer, the main log file for the Medulla agent is:

C:\Program Files\Medulla\var\log\xmpp-agent-machine.log

This file allows you to verify, in particular:

  • whether the agent started up;
  • connection attempts;
  • registration operations;
  • agent reconfigurations;
  • the presence of recurring errors.

When a computer appears offline in Medulla, this log is one of the first things to analyze.


3. Check whether the agent is reconfiguring itself in a loop

The Medulla operations guide indicates that a computer that is no longer reporting status correctly may be affected by repeated agent reconfigurations.

In the log:

C:\Program Files\Medulla\var\log\xmpp-agent-machine.log

look for the message:

We start a reconfiguration of the medulla agent

To count the number of occurrences on Windows:

find /c "We start a reconfiguration of the medulla agent" "C:\Program Files\Medulla\var\log\xmpp-agent-machine.log"

If this message appears a large number of times in the same file—for example, several dozen times—this may indicate that the agent is reconfiguring itself abnormally.

Caution:
A one-time restart or reconfiguration of the agent is not necessarily abnormal. It is primarily the frequent repetition of this behavior that should be analyzed.


4. Restart the Medulla agent

If the logs indicate abnormal agent behavior, the operations guide recommends restarting the agent to verify whether its operation stabilizes.

On a Windows computer:

sc stop medullaagent

Then stop any active Python processes related to the agent, if necessary:

taskkill /F /IM python.exe

Then restart the agent:

sc start medullaagent

Monitor the file:

C:\Program Files\Medulla\var\log\xmpp-agent-machine.log

to verify that the agent restarts normally and that the computer reconnects to Medulla.

Caution:
The command taskkill /F /IM python.exe terminates all python.exe processes on the computer. It should only be used if you have verified that no other important applications are using Python on the affected workstation.


5. Check the computer’s XMPP account

Medulla uses XMPP to communicate with its agents.

In some documented cases, when an agent is stuck in a reconfiguration loop, it may be necessary to delete its old ejabberd registration before allowing the agent to re-register.

From the ejabberd server:

ejabberdctl unregister ACCOUNT_NAME pulse

Then restart the Medulla agent on the computer:

sc start medullaagent

Caution:
Deleting an ejabberd account causes its XMPP registration to be deleted. This action should only be performed when a registration or reconfiguration issue has been clearly identified.


6. Check the known status of the computer in Medulla

The computer’s last known status can be viewed in the uptime_machine table.

SELECT *
FROM uptime_machine
WHERE hostname = "MACHINE_NAME"
ORDER BY id DESC
LIMIT 1\G;

The status field indicates the recorded status:

  • 1: the computer is considered online;
  • 0: the computer is considered powered off or offline.

The updowntime field indicates how long the computer has been in this state.

This information allows you, in particular, to compare the status displayed in Medulla with the computer’s actual status.


7. Check whether the computer is restarting or reconnecting repeatedly

It is also possible to search the xmppmaster database for computers that repeatedly report themselves as available over a short period of time.

The operations guide provides the following query:

SELECT COUNT(*) AS restartmach, hostname
FROM xmppmaster.uptime_machine
WHERE status = 1
  AND updowntime < 100
  AND date > (NOW() - INTERVAL 1 HOUR)
GROUP BY hostname
HAVING restartmach > 2;

This query identifies computers with multiple rapid status changes that may exhibit abnormal behavior.

Note:
The operations guide associates this query with the search for computers that may be restarting or reconnecting in a loop.


8. Verify the computer’s network information

An inconsistency between the computer’s information and its network interfaces may prevent it from registering correctly.

To view the known network information in Medulla:

SELECT
    m.hostname,
    m.ippublic,
    m.ip_xmpp,
    m.macaddress AS machine_mac,
    n.ipaddress AS network_ip,
    n.broadcast,
    n.gateway,
    n.mask,
    n.mac AS network_mac
FROM xmppmaster.machines AS m
INNER JOIN xmppmaster.network AS n
    ON n.machines_id = m.id
WHERE m.hostname = 'MACHINE_NAME';

In particular, verify the consistency of:

  • the hostname;
  • the IP address;
  • the MAC address;
  • network interface information.

9. Inconsistency between machines and network

In some cases, registering a computer may fail with messages similar to:

incoherence between machines and network tables

Interface missing on machine

Registration incomplete for machine

Several causes are listed in the operations guide:

  • network interface blacklisted;
  • duplicate MAC address;
  • network table entry without a corresponding computer.

To search for orphaned network entries:

SELECT *
FROM xmppmaster.network
WHERE machines_id NOT IN (
    SELECT id FROM xmppmaster.machines
);

Caution:
An anomaly in the machines or network tables relates more to a registration issue than to a simple offline status. Correcting these inconsistencies is covered in the FAQ dedicated to computer registration.


10. Verify relay server assignment

A computer must have a properly assigned relay server for certain Medulla operations.

The network information obtained previously can be used to verify whether the computer is on the expected network.

The groupdeploy field in the machines table contains the assignment information for the relay used for deployment.

If the computer also displays errors such as:

ABORT RELAY DOWN
ABORT INFO RELAY MISSING

the diagnosis should be continued at the relay server level and regarding its assignment.


11. Do not confuse an offline computer with a registration issue

Two different situations can produce similar symptoms:

Situation Interpretation
The computer was previously online but is no longer communicating Focus your troubleshooting on the agent, connectivity, and XMPP
The computer has never appeared correctly in Medulla Check its registration
The computer appears and then disappears periodically Check for repeated agent restarts or reconfigurations
The logs contain Interface missing or Registration incomplete Check for registration or network interface inconsistencies

12. Quick diagnostic procedure

When a computer appears offline in Medulla, perform the following checks in this order:

  1. verify that the computer is powered on and accessible;
  2. check the xmpp-agent-machine.log;
  3. check whether the agent is repeatedly reconfiguring itself;
  4. check and, if necessary, restart the MedullaAgent service;
  5. check the computer’s latest status in uptime_machine;
  6. verify the consistency of its network information;
  7. if the logs indicate a registration issue, proceed with the corresponding procedure.

13. Diagnostic table

Symptom Check Action
Computer is physically turned off Check the power supply and network connection Bring the computer back online
Agent stopped Check MedullaAgent Restart the service
Repeated reconfigurations Search for We start a reconfiguration... Diagnose the agent and its registration
Computer considered offline by Medulla Check uptime_machine Compare with the actual status of the workstation
Missing interface Check machines and network Continue the registration diagnosis
Duplicate MAC address Check network information Identify the duplicate computer
Incorrect or unavailable relay Check the relay assignment Diagnose the relay server

14. Information to collect if the problem persists

Before continuing with troubleshooting or opening a support ticket, gather the following:

  • the computer name;
  • its IP address;
  • the date and time when it first appeared offline;
  • the xmpp-agent-machine.log file;
  • the status of the MedullaAgent service;
  • the latest status available in uptime_machine;
  • any Interface missing or Registration incomplete messages;
  • the relay server associated with the computer if the issue also affects deployments.

Recommendation:
When a computer appears offline, start by determining whether the problem stems from the computer itself, the Medulla agent, or its registration. Checking the xmpp-agent-machine.log file and the latest known status in uptime_machine usually helps narrow down the diagnosis quickly.