A Basechain node with all historical state from the genesis block onwards takes up around 500GB of disk space, it's not practical to sync a new node from the genesis block (it would take forever). New nodes can be spun up using one of the jump start archives that we provide, these contain recent state, so a node can sync up with the chain in a few hours.
We provide two kinds of jump starts, a full size one, and a minimal one. The full size jump start contains a lot of historical data, you would generally only need it if you wanted to index all the historical state. If you just want to spin up a new validator node, or a sentry node, then you're better off using the minimal jump start, which only contains recent state.
Follow the instructions below to launch a Basechain node from a jump start archive.
Hardware Requirements
The minimum hardware requirements are as follows: 4 CPU, 32 GB RAM, NVMe SSD. Example: AWS i3.xlarge.
Download the Latest Basechain Stable Build
wget https://downloads.loomx.io/loom/linux/stable/basechain
chmod +x basechain
Download the Jump Start Archive
There are two available locations where you can download it from, USA or Singapore. Choose the location closest to you. The Singapore location is automatically replicated from the USA location so it might be a while before it becomes available when we update the archive.
In each location, there are two different archives. One is full and another is the smaller image that has no historical app state.
United States
- Minimal jump start:
- Archive: http://loom-share.s3-website.us-east-2.amazonaws.com/plasma-jump-start-min.tar.gz
- MD5: http://loom-share.s3-website.us-east-2.amazonaws.com/plasma-jump-start-min.tar.gz.md5sum
- [OPTIONAL] Receipts DB (not necessary for most new nodes)
- Full jump start:
Singapore
- Minimal jump start:
- Archive: http://loom-share-sg.s3-website.ap-southeast-1.amazonaws.com/plasma-jump-start-min.tar.gz
- MD5: http://loom-share-sg.s3-website.ap-southeast-1.amazonaws.com/plasma-jump-start-min.tar.gz.md5sum
- [OPTIONAL] Receipts DB (not necessary for most new nodes)
- Full jump start:
Extract the Archive
tar xfvz plasma-jump-start.tar.gz
Generate the Unique Config
mkdir blank
cd blank
cp ../loom-jump-start/loom.yml .
../basechain init
cp chaindata/config/{priv_validator.json,node_key.json} ../loom-jump-start/chaindata/config/
cd ../loom-jump-start
Start the Node
../basechain run --persistent-peers tcp://[email protected]:46656
Finally
The non-validator node will now sync with the validator nodes. Status can be queried at http://localhost:46658/rpc/status
Run Loom as a service on Ubuntu 16.04 (Optional)
Create /etc/systemd/system/loom.service
[Unit]
Description=Loom
After=network.target
[Service]
Type=simple
User=<user that runs loom>
WorkingDirectory=<working directory of loom>
ExecStart=<path to loom>/basechain run --persistent-peers tcp://[email protected]:46656
Restart=always
RestartSec=2
StartLimitInterval=0
LimitNOFILE=500000
StandardOutput=syslog
StandardError=syslog
[Install]
WantedBy=multi-user.target
Reload the Config
systemctl daemon-reload
Run Loom on Boot
systemctl enable loom.service
To start/stop/restart the Service
systemctl (start/stop/restart) loom.service
Inspect the Logs (Method 1)
tail -f /var/log/syslog
Inspect the Logs (method 2)
sudo journalctl -u loom.service -f
The Open Files Limit
When using systemd, the open files limit is taken care of by the LimitNOFILE directive. It will not, however, be applied when running loom manually.
The system-wide Limit
Add the following lines to the /etc/sysctl.conf file:
fs.file-max=500000
To activate it immediately:
sudo sysctl -p
To check whether it is active:
cat /proc/sys/fs/file-max
For a Specific User
Add the following lines to the /etc/security/limits.conf file:
username hard nofile 500000
username soft nofile 500000
Note: The user needs to re-login to have it active
To check whether it is active:
# check soft limit
ulimit -Sn
# check hard limit
ulimit -Hn