# Troubleshooting

> Diagnose a collector that is not running or a device that is not reporting.

Most collection problems fall into two groups: the agent is not running, or the agent is running but one device is failing. Work out which before going further, because the two have entirely different fixes.

## VSIClient will not stay running

If the scheduled task starts and then returns to `Ready` almost immediately, the agent is exiting early.

- Run the client manually in a command prompt as the service user and watch for errors on screen. This surfaces problems that the scheduler swallows.
- If no errors appear, close it and try starting the task again from Task Scheduler.
- If it still will not stay running, open the **History** tab for the task in Task Scheduler and look for error messages or failure codes.
- Check `\vsi\vsi_collection_process.log` for errors written during startup.

## A single device is not collecting

When the agent is healthy but one device produces no data, the collector's own log will say why.

1. Go to `visual1intelligence.com/user/settings/uploads`.
2. Find the upload entry for the device in question.
3. Download the corresponding ZIP file.
4. Open it and locate the file named `collection.log`.
5. Read the log for the failure.

Failures in that log almost always fall into three categories:

| What the log shows | Usual cause | Where to look |
| --- | --- | --- |
| Authentication errors | The credential is wrong, expired, or lacks the required role on the device | Confirm the stored password, then check the required role for that device type |
| Network or connection problems | The collection machine cannot reach the device on the required port | Confirm the port and that a firewall or ACL is not blocking it |
| Device-specific failures | A prerequisite is missing, or a required parameter was not supplied | Check the device's entry for prerequisites such as a vendor CLI, SDK, or serial number |

## Confirming a credential

To check the password Visual One has stored for a device, run this as the service user, with the device ID from the collection devices page:

```
keyring get <DEVICE_ID> password
```

## Reproducing a failure directly

Rather than waiting for the next scheduled run, reproduce the failure immediately with a manual collection. It runs independently of the scheduled agent and writes to a path you choose:

```
vsiclient --collector NETAPP_CLUSTER --ip netapp01.example.com --user svc-visualone --password EXAMPLE_PASSWORD --output_path C:\vsi\test
```

Remember that parameters differ by collector. A device needing a serial number, API token, or CLI path will fail without it even when the credential is correct.

## A collection runs but takes a very long time

Some collections are simply long. VMware and vSAN collections through vCenter can take several hours in large environments, which is expected rather than a fault. Check the upload timestamp before treating a slow collection as a failure.

## Related pages

[VSIClient Administration](https://visualoneintelligence.com/docs/vsiclient-administration/) for agent operations, and [Supported Devices](https://visualoneintelligence.com/docs/supported-devices-overview/) for the roles, ports, and prerequisites of each device type.

---

Source: https://visualoneintelligence.com/docs/troubleshooting/