This guide provides instructions for upgrading and troubleshooting common issues in Autonomys, based on feedback from Discord support channels.
How to Upgrade from Gemini/Taurus to Mainnet?
- Advanced CLI/Docker Users: Delete existing node database and farmer data. Ensure you’re using executables labeled “mainnet” and have specified “mainnet” as the chain.
- Space Acres Users: The upgrade is automatic, which will remove old node data and recreate your farms.
Where can I find the correct version of the software?
Is my Advanced CLI/Docker Configuration Correct?
This cannot be stressed enough: Most of the time the default parameters are what you want. Many times we are seeing old parameters that are outdated or parameters that are set incorrectly. For example, --sync full and --pot-external-entropy used to be required, but no longer is. Please refer to the documentation to make sure you have the correct parameters.
Why Is My Piece Cache Sync Taking So Long?
Autonomys relies on a peer-to-peer network for piece cache downloads. To improve sync times:
- Ports: Ensure ports 30333, 30433, and 30533 (for Advanced CLI) are open. For the piece cache, port 30533 is critical on Advanced CLI. You should ideally see 40 peers in node’s logs. Refer to the port configuration guide for more details.
- Connection Limits: Increase peer connections if needed, but beware of potential strain on your router, especially if others are using the internet.
- Advanced CLI Options : Increase peer connections
- Space Acres: Enable the “Faster Networking” option in advanced configuration.
For additional information about the piece cache, please see the Piece Cache Guide.
Why Is My GPU Not Always Busy?
- Data Bottleneck: GPUs require data ready for plotting. Check that ports are properly forwarded, especially port 30533 on Advanced CLI. For small networks, this is even more essential.
- Piece Cache: By default, the piece cache is 1% of your allocated space. If your allocated space is 4 TB, your cache will be 40 GB. Currently, in order to have a large enough piece cache to store the entire cache, you should have a farm size of around 12.5 TiB. In farming cluster setups, ensure your cache size is adequate relative to the archived history.
- Cluster Bandwidth: Farming clusters pass substantial data between NATS servers and components, potentially requiring bandwidth greater than 2.5 Gbps. Consider upgrading to 10 Gbps or greater if necessary.
Why Am I Replotting So Much?
In Phase 1 (until Wednesday the 13th), the network seeds data at 0.7 to 1.0 GiB per hour. Each time the history size doubles, 50% of sectors plotted up to that point will expire and need replotting. This process will slow down over time as it will take longer to double each time. If you want to avoid excessive replotting, consider waiting until Phase 2.
Why Am I Not Syncing or Staying in Sync?
Ensure ports are forwarded, especially if your peer count is below 38. In very rare cases, you can delete the node data with the app stopped, then restart. Snap sync should re-sync in 10-20 minutes.
Also note that the --sync full parameter is no longer required. Using it will make your sync take much longer.
Do I Need to Use the New “su” Address?
You can use either the new “su” address or your existing “st” address. SubWallet allows multiple address types under the same seed. For additional information, please see the Substrate Wallet Guide.
Unknown Variant V1 Error
If you get a message like this: ERROR {farm_index=0}: subspace_farmer::commands::farm: Farm creation failed error=Single disk farm I/O error: unknown variant v1, expected v0, this is caused by not removing your old testnet plots. Stop the farmer, delete the old testnet plots, and restart the farmer.