Disk Usage Optimization
Blockchain database tends to grow over time, depending e.g. on block speed and transaction amount. For BlockX, we are talking about close to 100GB of disk usage in first two weeks.
There are few configurations that can be done to reduce the required disk usage quite significantly. Some of these changes take full effect only when you do the configuration and start syncing from start with them in use.
If you do not need to query transactions from the specific node, you can disable indexing. On
indexer = "null"
If you do this on already synced node, the collected index is not purged automatically, you need to delete it manually. The index is located under the database directory with name
I believe this was disabled by default on BlockX, but listing it in any case here. On
snapshot-interval = 0
Note that if state-sync was enabled on the network and working properly, it would allow one to sync a new node in few minutes. But this node would not have the history.
By default every 500th state, and the last 100 states are kept. This consumes a lot of disk space on long run, and can be optimized with following custom configuration:
pruning = "custom"
pruning-keep-recent = "100"
pruning-keep-every = "0"
pruning-interval = "10"
pruning-keep-recent = "0"might sound tempting, but this will risk database corruption if the
blockxdis killed for any reason. Thus, it is recommended to keep the few latest states.
By default the logging level is set to
info, and this produces a lot of logs. This log level might be good when starting up to see that the node starts syncing properly. However, after you see the syncing is going smoothly, you can lower the log level to
config.tomlset the following
log_level = "warn"
Also ensure your log rotation is configured properly.
Below is the disk usage after two weeks of BlockX Arsia Mons testnet. The default configuration results in disk usage of 90GB.
This optimized configuration has reduced the disk usage to 17 GB.