Provider maintenance on your pod
A provider can announce downtime for the machine your pod runs on. The pod page then shows a yellow Maintenance Notice banner, and you get an e-mail if you were already renting.
Scheduled maintenance
The banner shows the window. It is at most 6 hours and announced at least 24 hours ahead (48 hours for one over 60 minutes).
What happens
The machine can go offline during the window. You are not charged for the window. After it, your pod keeps running if its container is still running. If the machine reports within three hours that it is not, Lium reboots a managed container pod once; on a bare-metal rental, coordinate recovery with your provider.
What survives
Files in /root and an attached Volume at /mnt. In a reboot, running jobs, memory and anything else are lost; reinstall packages you added outside /root.
What to do
- Before: save your work in
/rootor download it withlium scp. Stop or checkpoint long jobs. - After: log in and continue. If a managed container pod isn't back an hour after the window, reboot it from the pod page. On a bare-metal rental, coordinate recovery with your provider.
Permanent removal
The banner shows the removal time, announced at least 48 hours ahead.
The machine stays rentable until then. Its listing shows Removed at with the time (removal_at in the API), and if you rent it you get an e-mail with the time right away.
What happens
Your pod is deleted shortly after that time and can't be restored. Billing stops then.
What survives
Your Volume at /mnt and completed backups. On a managed container pod, everything else is deleted. A bare-metal rental is not wiped by Lium; arrange removal of your data with your provider.
What to do
Before the removal time, download what you need, copy it to a Volume, or run a backup. Then rent a new pod and attach the Volume or restore the backup.
If something went wrong
If your pod came back broken or stayed down much longer than the window, ask in Discord support with the pod name and the times from the banner.