Due to some performance issues with the node program of Conflux v2.2.0（https://github.com/Conflux-Chain/conflux-rust/releases） before December 6, 2022, it may not be able to catch up with the latest block or OOM issues.
It is recommended that all developers with nodes in operation do the following：
Option 1:
Program changes the configuration file:
Change the logger::level in the file pos_config/pos_config.yaml from INFO to WARN
and add:
storage:
prune_window: 10000000
Restart the node after you add it.
Option 2:
Directly replace the file link with the official modified file.file link
Note: How to upgrade nodes without being forcefully retired
It is recommended to notify the node to stop participating in the election before upgrading. After the node status changes to non-participating, you can shut down the node and upgrade it safely.
The node will automatically rejoin the PoS election after restarting. Specific steps are as follows:
- Run the command
./conflux rpc local pos stop_electionto inform the node to stop participating in the election. The node will not end its participation immediately, and the whole process may take several hours. The command returns theblock numberwhere the node will stop participating in the election. - Repeat this command at regular intervals to check the status of the node. If the command returns
null, then the node has stopped participation in the election. It can be shut down at this time. - Restart the node after the upgrade is completed. The node will automatically re-participate in voting. (Restarting a node will automatically re-execute from the latest checkpoint to the newest block. It may take a long time (several hours), but could be very fast as well)