Conflux v2.1.0 Network Hardfork Upgrade Announcement

Conflux Network will be upgraded when the Epoch Number reaches 56800000 (estimated on Oct. 18th, 2022 (GMT+8)). All nodes need to be upgraded before the time mentioned above. Most CIPs (Conflux Improvement Proposals) will be enabled when the Block Number reaches 133800000 (estimated on Oct. 25th, 2022 (GMT+8)), and CIP-99 will be enabled when the PoS Block Number reaches 330000 (estimated on Oct. 25th, 2022 (GMT+8)).

The new fullnode program’s download link:

Upgrade Contents

Incompatible Changes

This Hardfork will activate 5 different CIPs (Conflux Improvement Proposals) including CIP-94, CIP-97, CIP-98, CIP-99, and CIP-105.

Configuration Improvement

  • Add Fullstate feature, configure enable_single_mpt_storage to enable storing all history states. Set single_mpt_space = "evm" to only store eSpace states. Note that you’ll need to resync an archive node to reconstruct all the history states.

Storage Improvement

  • Allow snapshots to be removed during recovery. This reduces the disk space requirement for full node sync or recovery.

RPC Improvement

  • Support eth_subscribe and eth_unsubscribe in eSpace RPCs.
  • Add a RPC method cfx_getParamsFromVote to return the currently used value of the voted parameters.
  • Return null for getting skipped transactions and receipts. Whether they will be returned was non-deterministic before, but now they are ensured to be null.
  • When querying logs, use LatestState instead of the to epoch number when the to epoch number is greater than LatestState.

Transaction Pool Improvement

  • Fix an issue that may fail to recycle packed but skipped transactions.

What will happen if I do not upgrade the nodes or refuse to upgrade?

If the Conflux node is not upgraded to the latest version, the node will be incompatible with the upgraded Conflux protocol and cannot interact with the latest Conflux Network as expected (functions like synchronizing blocks, mining, etc might be disabled).

To node operators and miners

  • If you upgrade your node before the Epoch Number reaches 56800000 (estimated on Oct. 18th, 2022 (GMT+8)), you need to replace conflux.exe with the new one in Conflux v2.1.0 under the run directory, then restart the node.
  • If you upgrade your node after the Epoch Number reaches 56800000 (estimated on Oct. 18th, 2022 (GMT+8)), you need to delete all data and restart the node.
  • If you start a new node, you can download the latest version of Conflux v2.1.0 and run it.

Note: using enable_single_mpt_storage to enable Fullstate feature will increase the disk space usage.

Reference: up to Epoch Number 55160017, additional_maintained_snapshot_count=0

variable setting ext4 xfs
Enable fullstate 970G 798G
Disable fullstate 556G 418G

Note: with this feature enabled, some disk space usage can be saved by not setting additional snapshot.

Note: Influence of CIP-99

According to CIP-99 ( ), after this Hardfork, more not-voting terms will be allowed before we force-retire a node, and make the unlock period of a retiring node shorter to allow the node to rejoin the PoS voting faster.

To ordinary users:

The upgrade will not affect the normal use of the Conflux Network. However, we advise you to avoid using the network due to instability and operational delay during the upgrade. There is a possibility that your asset on Conflux Network will be lost.

According to CIP-94 ( ), after this Hardfork, users can use on-chain DAO voting to decide and update reward parameters without Hardfork.

Thanks for your support!

A sincere thank you to the Conflux community and all the developers in the Conflux ecosystem. Thank you for your hard work, thoughts, and contributions. To a better future!