3. Syncing DB

Simple algorithm to distinguish between snapshot types

Using answers for these questions you can easily understand which snapshot type you will need

Validator Nodes

This document introduces another centralized fast state syncing method using rclone. Please use it with caution. This guide is mainly used for a newly started node to catch up with the blockchain faster. Otherwise, the blockchain syncing may take weeks from genesis block.

Rclone db snapshot is synced with blockchain frequently. However, there maybe a potential race condition when the rclone may fail due to our nodes were updating the db files at the same time. In this case, just re-run the rclone command to re-sync again.

1. Installing Rclone

For installing Rclone, please follow the instructions at https://rclone.org.

TL;DR: on a Linux system, you may run the following command:

curl https://rclone.org/install.sh | sudo bash

Make sure your rclone version is above v1.53.2 . you can check version by rclone version

2. Configuring Rclone

To check the location of the rclone.conffile run:

rclone config file

The rclone.conf file is usually located at ~/.config/rclone/rclone.conf .

Now run the following command to create the rclone.conf file:

cat<<-EOF > ~/.config/rclone/rclone.conf
[snap]
type = webdav
vendor = other
user = snap
pass = ufbTDtK0fENuutwuDOHae57xT8URsZVIcdotK30T5A
EOF

2.1 BEWARE: Understanding the `rclone sync`

The most important parts in bold:

Sync the source to the destination, changing the destination only. Doesn't transfer files that are identical on source and destination, testing by size and modification time or MD5SUM. Destination is updated to match source, including deleting files if necessary (except duplicate objects). Link to the full rclone documentation.

sync command was chosen because you need to have your blockchain data folder to be identical with the snapshot data

As user you need to carefully get the full path to the folder and as destination. If you have some important data in the destination, please move it somewhere else.

Usually data dir folder is `./` which means folder in which harmony is running

You can double check this via:

ls -l | grep 'harmony_db_'

Expected result id is name with shard number on the end, e.x. harmony_db_0.

Q: How to get the full path to the your blockchain data folder if it isn't default?

1. Check how do you start harmony via ps -ef | grep harmony and check for the --datadir option, if you don't have it go to the following step, else to the 3rd

2. Check the harmony.conf for DataDir option or check your grep DataDir harmony.conf

3. Go to the folder from the cmd option if have it or to folder from config if you don't have cmd option, do the following:

cd your_data_dir # folder from the cmd line or config
ls -l # it will give the shard folder name like harmony_db_0/
pwd # now you have full path to your data dir

3. Shard 0 validator - Snap DB sync

SnapDB is a type of DB that contain a snapshot of the state at a given block. It doesn't have any block history and hence not suitable for RPC node.

Below is the command to sync the Snap DB for the shard 0. It is around 100 Gb as of October 2023:

Important: Since this can cause data loss, test first with the --dry-run or the --interactive/-i flag.

YOUR_DATA_DIR='./' # This is the default installation, change it you use something else
rclone -P -L --webdav-url 'http://snapdb.s0.t.hmny.io/webdav'  --checksum sync \
  snap: "${YOUR_DATA_DIR}harmony_db_0" --multi-thread-streams 4 \
  --transfers=32 --fast-list --verbose

Small explanation for flags used to save your time with rclone manual:

# --webdav-url - source of the data and URL of http host to connect to
# YOUR_DATA_DIR/harmony_db_0 - your destination
# -P - flag to show the progress
# -L - Follow symlinks and copy the pointed to item
# --checksum - Normally rclone will look at modification time and size of 
#  files to see if they are equal.  If you set this flag then rclone will 
#  check the file hash and size to determine if files are equal
# snap: - the way how do you tell which config is used 
# --multi-thread-streams 4 - When using multi thread downloads this 
#  sets the maximum number of streams to use
# --transfers=32 - The number of file transfers to run in parallel.
# --fast-list - load all file paths into memory at once instead of walking each directory one-by-one
#   It can sometimes be useful to set this to a smaller number if the remote is giving a lot of timeouts or bigger if you have lots of bandwidth and a fast remote.
# --verbose - With -v rclone will tell you about each file that is transferred and a small number of significant events.
# sync - make the origin and destination the same, after 1st run add the deltas

4. Full db sync for appropriate shard

Since v4.3.12 shard 1 doesn't need to download anymore shard 0 DB. Instead, the node will download automatically the blocks required for the node in shard 1 to work.

Each node will simply need to rclone its own DB.

4.1 Shard0 RPC Explorer node (non archival)

Important: Since this can cause data loss, test first with the --dry-run or the --interactive/-i flag.

YOUR_DATA_DIR='./' # This is the default installation, change it you use something else
rclone -P -L --webdav-url 'http://fulldb.s0.t.hmny.io/webdav' \
   --checksum sync snap: "${YOUR_DATA_DIR}harmony_db_0" --multi-thread-streams 4 \
   --transfers=32 --fast-list --verbose

4.2 Shard 1 validator

Important: Since this can cause data loss, test first with the --dry-run or the --interactive/-i flag.

YOUR_DATA_DIR='./' # This is the default installation, change it you use something else
rclone -P -L --webdav-url 'http://fulldb.s1.t.hmny.io/webdav' \
   --checksum sync snap: "${YOUR_DATA_DIR}harmony_db_1" --multi-thread-streams 4 \
   --transfers=32 --fast-list --verbose 

Archival snapshot for the Non-Validating/Explorer Nodes

As of 1st October 2023, the size for the shard 0 on mainnet is ~23TiB. Note that the bucket is currently not fully synced.

Please contact us at [email protected] if you need to access our archival db.

5. Testnet snapshots

If you want to quickly test your validator/RPC setup via the testnet network, please use the setup from 2. Configuring Rcloneand the info below.

5.1 Testnet shard 0

Important: Since this can cause data loss, test first with the --dry-run or the --interactive/-i flag.

YOUR_DATA_DIR='./' # This is the default installation, change it you use something else
rclone -P -L --webdav-url 'http://fulldb.s0.b.hmny.io/webdav' \
   --checksum sync snap: "${YOUR_DATA_DIR}harmony_db_0" --multi-thread-streams 4 \
   --transfers=32 --fast-list --verbose 

5.2 Testnet shard 1

Important: Since this can cause data loss, test first with the --dry-run or the --interactive/-i flag.

YOUR_DATA_DIR='./' # This is the default installation, change it you use something else
rclone -P -L --webdav-url 'http://fulldb.s1.b.hmny.io/webdav' \
   --checksum sync snap: "${YOUR_DATA_DIR}harmony_db_1" --multi-thread-streams 4 \
   --transfers=32 --fast-list --verbose