> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.itential.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.itential.com/_mcp/server.

# Start, stop, and restart Platform

> Managing Itential Platform service lifecycle

Use the information on this page to start, stop and restart Itential Platform. Information on handling a force quit is also provided.

Always restart the Platform after properties have been updated. Profile services and properties can be accessed via [Admin Essentials](/itential-platform/admin-essentials/overview).

## Service & Systemctl

There are two Linux-based management tools that provide a way to manage system services: `service` and `systemctl`. In the case of `service`, the command will redirect to `systemctl` on systemctl-enabled distributions. Both service and systemctl commands for starting, stopping and restarting services can be run on both CentOS and RHEL distributions. The systemctl commands can be run on Rocky Linux.

The current `systemd` service-name is `itential-platform`.

```bash
systemctl <command> itential-platform
```

Where `<command>` will be "stop", "start", "restart" and "status".

## Start

Run the following to start the Platform.

**service command**

```bash
sudo service pronghorn start
```

**systemctl command**

```bash
sudo systemctl start itential-platform
```

Use this command to start the Platform as a service. With this command, the server will continue to run after you end the session. You can also use this command to manually start the Platform from the home directory.

```bash
sudo -u pronghorn node server.js
```

This alternate command starts up a temporary node. When you end the session with this alternate command, the server will stop.

## Stop

Run the following to stop the Platform.

**service command**

```bash
sudo service pronghorn stop
```

**systemctl command**

```bash
sudo systemctl stop itential-platform
```

## Restart

Run the following to restart the Platform.

**service command**

```bash
sudo service pronghorn restart
```

**systemctl command**

```bash
sudo systemctl restart itential-platform
```

This command starts the Platform as a service. Refer to the Start section above for more information.

## Verify Stop/Start Status

To verify the Platform started/stopped successfully, use the system log file.

**less command line utility**

```bash
sudo less /var/log/pronghorn/pronghorn.log
sudo tail -f /var/log/pronghorn/pronghorn.log
```

**systemd utility**

```bash
sudo journalctl -xe
sudo journalctl -fu itential-platform
```

In a VM, these commands may be run from the root directory.

## Force Quit

Itential highly recommends that customers **never** force quit the Platform, or use a `kill -9` command on the Platform process.

### How to Handle a Force Quit if the Platform is Never Restarted

In situations where customers do force quit the Platform and that Platform instance is never restarted, some additional measures are needed. A course of action to clean up Redis data is required and you must know the `serverId` of the Platform instance killed.

### Find the serverId

Use this Redis command to find the `serverId` if you do not have it:

```bash
smembers iap:servers
```

### Retrieve health keys

In the Redis CLI, run the following commands to retrieve the health keys:

```bash
srem iap:servers $IAP_ID
keys health:*:$IAP_ID
```

### Delete the keys

Next, run each health key through the delete command:

```bash
del $KEY1
```

Alternately, to avoid running `del` multiple times, use a space-separated list. Each key is the result of the `keys` command run previously:

```bash
del $KEY1 $KEY2 $KEY3 $KEY4
```

### Consequences if No Cleanup is Performed

If a cleanup is not performed, the ramifications are minimal; however, if the Platform is stopped via force quit and never restarted, the Redis memory becomes bloated over time and leads to an "out-of-memory" condition. To avoid this altogether, run the cleanup steps above or just simply restart the Platform so it can self-clean.