Hello Access node operators,
In the upcoming Flow mainnet spork, execution data syncing will be enabled by default on all access nodes after the upgrade. The execution data sync lets the access node download execution data, such as transactions and events, directly from the execution node. This data can then be used to serve the ExecutionDataAPI, including the new event streaming API.
The execution data will increase the disk usage on the Access node, and hence, please ensure you have at least 750GB of disk space for the Access node before the next mainnet spork, which will be in around six weeks’ time. Hardware requirements can be found here.
You can monitor the disk usage percentage by monitoring the Prometheus metric:
(1 - (node_filesystem_avail_bytes{mountpoint="/var/flow"} / node_filesystem_size_bytes{mountpoint="/var/flow"}))
The execution data directory can be set with the flag: --execution-data-dir=/data/execution_data
By default, execution data sync will be enabled. It can be disabled with the following flag,
--execution-data-sync-enabled=false; however, we strongly encourage you to keep the flag enabled. Access nodes share the execution data among themselves on the peer-to-peer, and to enable faster data dissemination, all access nodes should allow execution data sharing.
Thank you,
Flow Team