Backup/restore node instance
Overview
This guide goes into detail on how to back up and restore a ZChains node instance. It covers the base folders and what they contain, as well as which files are critical for performing a successful backup and restore.
Base folders
ZChains leverages LevelDB as its storage engine. When starting a ZChains node, the following sub-folders are created in the specified working directory:
blockchain - Stores the blockchain data
trie - Stores the Merkle tries (world state data)
keystore - Stores private keys for the client. This includes the libp2p private key and the sealing/validator private key
consensus - Stores any consensus information that the client might need while working. For now, it stores the node's private validator key
It is critical for these folders to be preserved in order for the ZChains instance to run smoothly.
Create backup from a running node and restore for new node
This section guides you through creating archive data of the blockchain in a running node and restoring it in another instance.
Backup
backup command fetches blocks from a running node by gRPC and generates an archive file. If --from and --to are not given in the command, this command will fetch blocks from genesis to latest.
Restore
A server imports blocks from an archive at the start when starting with --restore flag. Please make sure that there is a key for new node. To find out more about importing or generating keys, visit the Secret Managers section.
Back up/Restore Whole data
This section guides you through backup the data including state data and key and restoring into the new instance.
Step 1: Stop the running client
Since the ZChains uses LevelDB for data storage, the node needs to be stopped for the duration of the backup, as LevelDB doesn't allow for concurrent access to its database files.
Additionally, the ZChains also does data flushing on close.
The first step involves stopping the running client (either through a service manager or some other mechanism that sends a SIGINT signal to the process), so it can trigger 2 events while gracefully shutting down:
Running data flush to disk
Release of the DB files lock by LevelDB
Step 2: Backup the directory
Now that the client is not running, the data directory can be backed up to another medium. Keep in mind that the files with a .key extension contain the private key data that can be used to impersonate the current node, and they should never be shared with a third/unknown party.
:::info Please back up and restore the generated genesis file manually, so the restored node is fully operational. :::
Restore
Step 1: Stop the running client
If any instance of the ZChains is running, it needs to be stopped in order for step 2 to be successful.
Step 2: Copy the backed up data directory to the desired folder
Once the client is not running, the data directory which was previously backed up can be copied over to the desired folder. Additionally, restore the previously copied genesis file.
Step 3: Run the ZChains client while specifying the correct data directory
In order for the Polygon Edge to use the restored data directory, at launch, the user needs to specify the path to the data directory. Please consult the CLI Commands section on information regarding the data-dir flag.
Last updated