Bitcoin node maintenance checklist

A node that has finished syncing mostly runs itself. Maintenance is confirming that, keeping disk free, and applying verified upgrades. Commands are from the commands library; install details are in setup.

What healthy looks like

$ bitcoin-cli getblockchaininfo | grep -E 'blocks|headers|verificationprogress|initialblockdownload'
  "blocks": 969640,
  "headers": 969640,
  "verificationprogress": 0.99999,
  "initialblockdownload": false,
$ bitcoin-cli getconnectioncount
10
$ df -h /var/lib/bitcoind | tail -1
/dev/nvme0n1p2  916G  812G   58G  93% /data

blocks equals headers, initialblockdownload is false, you have peers, and the service is up. The last line above is a warning: 93% full on a growing chain means act this month (disk recovery). Still syncing? Use sync stuck instead. You can compare against a healthy and an unhealthy node in the lab.

Weekly (five minutes)

• Service is up: systemctl status bitcoind
• blocks equals headers in getblockchaininfo
• getconnectioncount is above zero
• Disk has headroom: df -h on the datadir filesystem
• Read logs only if something is off: journalctl -u bitcoind -n 40

Monthly

• OS updates applied
• RPC still local-only: ss -tlnp | grep 8332 shows 127.0.0.1 (security basics)
• Your version vs the latest stable on bitcoincore.org: bitcoind --version
• VPS only: bandwidth and volume size against provider limits
• Router still forwards 8333, if you want inbound peers

Upgrading Bitcoin Core

1. Read the release notes for config changes.
2. Download from bitcoincore.org and verify the signatures (steps in setup).
3. Stop cleanly and swap the binaries in the path your unit uses.
4. Start, confirm getblockchaininfo answers, and watch the log for a few minutes.

$ bitcoind --version | head -1
$ sudo systemctl stop bitcoind
# install the verified binaries (see /setup/)
$ sudo systemctl start bitcoind
$ bitcoin-cli getblockchaininfo | grep -E 'blocks|headers'
$ journalctl -u bitcoind -n 50 --no-pager

Right after a restart bitcoin-cli may answer error code: -28 while Core loads its block index. Wait a minute and retry.

Leave it alone

A healthy node does not need daily restarts or tuning. Avoid exposing RPC "for a dashboard", deleting anything under blocks/ or chainstate/, and third-party "optimizer" scripts. Adding an Electrum server is a RAM and disk decision; see Electrum server vs full node.

Home and VPS differences

At home, watch for a full root partition, a router change that drops 8333, and an unplanned power loss (a UPS helps; a hard power cut mid-write is the usual cause of a corrupted block index). On a VPS, watch provider maintenance emails, bandwidth and volume size. Reach a home node remotely over a VPN or SSH, never an open RPC port. Hardware that stays on 24/7 is on the hardware page; hosted options on VPS.

Logs

Most of debug.log is peer churn. Act on repeated disk errors, LevelDB complaints or crash loops. Logs contain peer IPs and paths, so redact before pasting them anywhere public.

Once a year, recheck whether your disk still has room for growth and whether pruned or full still fits your use (pruned vs full).

FAQ