Overview

• KV_SAVE persists all keys
• Data can be restored at startup with a command line argument
• Data can be restored at runtime with KV_LOAD
• Data is written as raw JSON, a future release will support a more space efficient format

Persisting Key Values

The KV_SAVE command persists key values.

The command saves all keys, you cannot specify particular keys to persist.

{
  "KV_SAVE":
  {
    "name":"stats"
  }
}

And later load at runtime with:

{
  "KV_LOAD":
  {
    "name":"stats"
  }
}

Loading

• Startup

  • --loadName will load data. The name must match a name used in KV_SAVE
  • --loadPath is used to read data from a path that's different from that in the server config
    • Only applies at startup, not with KV_LOAD

• Runtime

  • KV_LOAD offer the flexibility to load data at any time
  • KV_LOAD only read data from the persist path in the server config

Persistance Structure

Data is saved with a name and a timestamp. The name is used when loading as a way to identify which data to restore.

The server's config file has a setting for the save path, in to which the data is written.

The structure is:

└── <name>
    └── <timestamp>
        ├── data
        └── md

Directory	Purpose
name	The name used in the save command
timestamp	Timestamp when the data was saved
data	Contains the data
md	Contains metadata

This structure allows you to save and restore specific data in a future release.

For example, if we save data using the name defaults we will have:

.
└── defaults
    └── 1701060607618993128
        ├── data
        └── md

If we send another save command with the same name:

└── defaults
    ├── 1701060607618993128
    │   ├── data
    │   └── md
    └── 1701061180908039731
        ├── data
        └── md

info

The current version does not allow you to load a specific timestamp, it always selects the most recent.

If you need to restore specific data, use a different name for each use of KV_SAVE.