-
Notifications
You must be signed in to change notification settings - Fork 41
Open
Description
We should add a dedicated section about pruning, probably under Maintenance section?
- How to configure it for consensus and paratime nodes.
- We have this currently under hardware requirements.
- Suggested values for the pruning.
- How to guarantee optimal disk usage:
- For now by pruning from the start.
- Later we will add the
oasis-node storage compactcommand there, possibly "offline pruning".
- Explain that with many versions, pruning is quite slow (especially for the runtime state) and that operators should consider (temporary) increasing BadgerDB block cache size.
- Especially when keeping months of the data, if possible. Pending some more experiments from Pruning of runtime state is too slow oasis-core#6334.
- How much extra free disk space operator should have available to guarantee compaction success.
Metadata
Metadata
Assignees
Labels
No labels