Sync snap errors

Support Incentivized Testnet
1

Good day, dear developers.
After ѕync snap function appeared, I decided to try it out. First impression: wow! Much less space, fast synchronization.
Disadvantages: at any restart of a node, whether it is an unplanned restart (sudden machine restart), or it is a planned restart of a node (for example, systemctl restart …), the node refuses to start and gives this error:

2025-05-27T16:41:17.320479Z INFO Domain: sc_sysinfo: :laptop: Operating system: linux
2025-05-27T16:41:17.320485Z INFO Domain: sc_sysinfo: :laptop: CPU architecture: x86_64
2025-05-27T16:41:17.320487Z INFO Domain: sc_sysinfo: :laptop: Target environment: gnu
2025-05-27T16:41:17.320489Z INFO Domain: sc_sysinfo: :laptop: CPU: AMD Ryzen 9 7950X 16-Core Processor
2025-05-27T16:41:17.320491Z INFO Domain: sc_sysinfo: :laptop: CPU cores: 16
2025-05-27T16:41:17.320493Z INFO Domain: sc_sysinfo: :laptop: Memory: 63475MB
2025-05-27T16:41:17.320496Z INFO Domain: sc_sysinfo: :laptop: Kernel: 6.8.0-59-generic
2025-05-27T16:41:17.320497Z INFO Domain: sc_sysinfo: :laptop: Linux distribution: Ubuntu 22.04.5 LTS
2025-05-27T16:41:17.320499Z INFO Domain: sc_sysinfo: :laptop: Virtual machine: no
2025-05-27T16:41:17.320513Z INFO Domain: sc_service::builder: :package: Highest known block at #1692729
2025-05-27T16:41:17.320616Z INFO Domain: sc_rpc_server: Running JSON-RPC server: addr=127.0.0.1:36697,
2025-05-27T16:41:17.320685Z ERROR Domain: domain_snap_sync: Snap sync can only work with genesis state.
2025-05-27T16:41:17.320690Z INFO Domain: domain_snap_sync: Wipe the DB and restart the application with --sync=full.
2025-05-27T16:41:17.320701Z ERROR Domain: sc_service::task_manager: Essential task domain-sync failed. Shutting down service.
2025-05-27T16:41:17.320788Z ERROR Domain: subspace_node::commands::run: Domain starter exited with an error error=Other: Essential task failed.
2025-05-27T16:41:17.320796Z ERROR sc_service::task_manager: Essential task domain failed. Shutting down service.
2025-05-27T16:41:17.431059Z WARN Consensus: sc_proof_of_time::source::gossip: Failed to send incoming message error=send failed because receiver is gone
2025-05-27T16:41:17.431084Z ERROR Consensus: sc_proof_of_time::source::gossip: Gossip engine has terminated
2025-05-27T16:41:17.431151Z ERROR Consensus: sc_service::task_manager: Essential task pot-gossip failed. Shutting down service.
Error: SubstrateService(Other(“Essential task failed.”))

After clearing the db folder and restarting the node, the synchronization starts from 0 (this is expected), but without errors.
Tried two versions of node:

  • taurus-2025-may-15;
  • taurus-2025-may-27;

PS. Node startup command
ExecStart=/usr/local/bin/subspace-node run --base-path /home/2tb/subspace-node --name Some-kind-of-name --chain taurus --sync snap --farmer --blocks-pruning archive-canonical --state-pruning archive-canonical --bootstrap-node /dns/bootstrap-0.taurus.ai3. farm/tcp/11105/p2p/12D3KooWDpPPWd7htCiSFaJq5WJPuX1iksBYZBCjnoeRberCmag6 --bootstrap-node /dns/bootstrap-1.taurus.ai3. farm/tcp/11106/p2p/12D3KooWCxMRLLXihZdMewPVtU2jPr5MdDFHcEnUCuZwvfX7bufB – --domain-id 0 --blocks-pruning archive-canonical --state-pruning 28800

1 Like
2

Your start command should avoid overriding the default pruning as per the latest docs unless you are running an archive node? You shouldn’t need the bootstrap overrides either. This should work for you:

  run
  --base-path /home/2tb/subspace-node
  --name Some-kind-of-name
  --chain taurus
  --sync snap
  --farmer
  --
  --domain-id 0

I’m not 100% sure but I would imagine that snap sync and archive-canonical are incompatible as you’re asking to snap sync (skip a bunch of historical blocks) but still be an archive node.

2 Likes