Home > Storage > PowerScale (Isilon) > Product Documentation > Cloud > APEX File Storage for AWS: Deployment Guide > Stopping OneFS node instances
OneFS nodes use one of the instance store volumes as the journal device. OneFS needs to save the content of the journal before stopping an instance since any data on the volume is lost when the instance is stopped. OneFS nodes use EBS volume as their root device and save the contents of the journal to the EBS volume during nodes’ shutdown.
When stopping an instance, while you can do so from the AWS Management Console directly, the best practice is to shut the instance down gracefully, by using the OneFS CLI command isi cluster shutdown --node-lnn=<node-lnn> or by using the OneFS WebUI. This also takes the instance to a ‘stopped’ state from the AWS perspective.
It is recommended not to force an instance to stop from the AWS Management Console using Force stop instance or using the AWS CLI with the 'force' option. Forcing instances to stop will not give the nodes an opportunity to save the journal and the nodes will be lost if they happen to restart on a new host hardware.
To prevent your instance from being accidentally stopped or terminated, you can enable stop protection or termination protection for the instance. See the AWS documentation Enable stop protection and Enable termination protection for details.
If the journal was not saved successfully when a node went down, the node will fail to boot while restoring the content from the journal. This may also happen when a node is restarted after it went down due to host hardware failures. Though it rarely happens, a system status check failure will indicate such problem with the AWS system on which the OneFS node was running. The only way to recover the lost storage space due to the OneFS node failing to boot up is by performing the node replacement.