For this release, we changed our contract from STATE to DVT so operators need to do the following steps to run your new operator nodes if you are still running an old operator node:
cd dvf
git pull && git checkout main
mv .env.example .env
# update .env with your configuration
sudo docker stop dvf-operator-1
sudo mv /data/operator /data/operator-backup
sudo mkdir -pv /data/operator/
sudo docker compose -f docker-compose-operator.yml up dvf_key_tool
# Save the public key, which will be used later. Go to SafeStake website and Join As Operator.
# After you register an Operator on the Safestake website, you will be shown your OPERATOR ID, which is the unique identifier you need to start with. You will need to update the OPERATOR ID to the .env file before running the operator service.
sudo docker compose -f docker-compose-operator.yml up --force-recreate -d operator
Change log
- Change the way of monitoring contract events to actively synchronizing events from the contract, and persist all events to the local database. Even if the node is down, it can be recovered after restart.
- Polling to handle all events. Failed events will remain local to increase fault tolerance
- Added initiator role for State2 testing
- Fix logics of VA add / delete
- Fix issue: rocksdb never exit
- Upgrade IP management in VA lifetime. Allow operators to update IP and reconnect. Allow non-responsive operators in registration
- Fix issue: “Stop validator” didn’t delete VA from memory, causing the VA still running
- Fix issue: When VA on OP1 stops, the DVF signer exits (and consensus exits), but handler is not destroyed. OP2 might keep running and sending messages to OP1 who will dispatch the messages as usual, and cause the thread to panic because the tokio receiver channel has been destroyed due to exit signal
- Fix issue: activating validator doesn’t work because it calls “add” validator, which results in a duplicated key error
- Fix issue: heavy logging messages. Add message handling delay to alleviate the issue