Validator operations (day-2)

This page is the operator's day-to-day cheat sheet: the commands you will run weekly or monthly once the validator is live. Everything assumes:

• You have your operator key in validator (keyring backend file).
• --node points at a healthy RPC (use a foundation endpoint or one of your own sentries).
• You are on mainnet; for testnet replace the chain id and RPC.

export VALADDR=$(safrochaind keys show validator --bech val -a --keyring-backend file)
export CHAIN_ID=safro-mainnet-1
export NODE=https://rpc.safrochain.network:443
export FEES="--gas auto --gas-adjustment 1.3 --fees 5000usaf"

Edit description / commission

safrochaind tx staking edit-validator \
  --moniker "my-safro-validator" \
  --website "https://example.com" \
  --identity "1234567890ABCDEF" \
  --details "Run by Example Co.: Africa-first" \
  --security-contact "[email protected]" \
  --commission-rate "0.05" \
  --from validator --keyring-backend file \
  --chain-id $CHAIN_ID --node $NODE $FEES --yes

Commission constraints:

• You can lower it any time.
• You can raise it once per commission-max-change-rate window (typically 24 h) and never above commission-max-rate.
• commission-max-rate and commission-max-change-rate are fixed at create-time and cannot be raised later.

Self-delegate more

safrochaind tx staking delegate $VALADDR 50000000usaf \
  --from validator --keyring-backend file \
  --chain-id $CHAIN_ID --node $NODE $FEES --yes

(50000000usaf = 50 SAF.)

Withdraw rewards

Rewards (commission + delegations) accrue automatically; you must withdraw them.

safrochaind tx distribution withdraw-rewards $VALADDR \
  --commission \
  --from validator --keyring-backend file \
  --chain-id $CHAIN_ID --node $NODE $FEES --yes

--commission also pulls the commission share. Without that flag you only pull rewards on your self-delegation.

Inspect what is pending first:

safrochaind query distribution rewards \
  $(safrochaind keys show validator -a) $VALADDR --node $NODE

Re-delegate (move stake to another validator)

safrochaind tx staking redelegate $VALADDR \
  addr_safrovaloper1otherValidatorAddressHere \
  10000000usaf \
  --from validator --keyring-backend file \
  --chain-id $CHAIN_ID --node $NODE $FEES --yes

A redelegation completes immediately but re-redelegating the same stake is rate-limited (one redelegation per validator pair per unbonding period).

Begin unbonding (graceful exit)

safrochaind tx staking unbond $VALADDR 100000000usaf \
  --from validator --keyring-backend file \
  --chain-id $CHAIN_ID --node $NODE $FEES --yes

The unbonded amount enters the unbonding queue and can still be slashed for offences committed before the unbond began. Until the unbonding period ends, the validator must keep signing.

Unjail after downtime

If the chain jails you for missing too many blocks, fix the cause then unjail:

safrochaind tx slashing unjail \
  --from validator --keyring-backend file \
  --chain-id $CHAIN_ID --node $NODE $FEES --yes

A validator still jailed error means you are unjailing too early: wait one more signed_blocks_window and retry. See Slashing & jail.

Software upgrades

Cosmos chains coordinate software upgrades through the x/upgrade module. Validators must run the new binary at exactly the upgrade height or the chain will halt.

The recommended pattern is Cosmovisor, which swaps binaries automatically on the upgrade height.

Install Cosmovisor

go install cosmossdk.io/tools/cosmovisor/cmd/cosmovisor@latest

# Cosmovisor expects this layout:
# $DAEMON_HOME/cosmovisor/
#     genesis/bin/safrochaind             # current binary
#     upgrades/<upgrade-name>/bin/safrochaind   # new binary, pre-staged

/etc/systemd/system/safrochaind.service:

[Unit]
Description=Safrochain via Cosmovisor
After=network-online.target

[Service]
User=safro
Environment=DAEMON_HOME=/home/safro/.safrochain
Environment=DAEMON_NAME=safrochaind
Environment=DAEMON_RESTART_AFTER_UPGRADE=true
Environment=DAEMON_ALLOW_DOWNLOAD_BINARIES=false
ExecStart=/home/safro/go/bin/cosmovisor run start
Restart=always
RestartSec=3
LimitNOFILE=65535

[Install]
WantedBy=multi-user.target

Set DAEMON_ALLOW_DOWNLOAD_BINARIES=false on the validator. Always pre-stage the upgrade binary yourself; never let the validator pull code from the internet.

Pre-stage the new binary

Hours (not minutes) before the upgrade height:

mkdir -p $HOME/.safrochain/cosmovisor/upgrades/v0.3.0/bin
# build or copy the v0.3.0 binary into:
cp ./safrochaind $HOME/.safrochain/cosmovisor/upgrades/v0.3.0/bin/safrochaind
chmod +x $HOME/.safrochain/cosmovisor/upgrades/v0.3.0/bin/safrochaind

# verify the version *the cosmovisor will run* matches the announcement
$HOME/.safrochain/cosmovisor/upgrades/v0.3.0/bin/safrochaind version

What happens at the upgrade height

1. The chain pauses one block before the upgrade height.
2. Cosmovisor swaps genesis/bin/safrochaind → upgrades/v0.3.0/bin/safrochaind.
3. safrochaind restarts and applies the migration handler.
4. The chain resumes at the upgrade height.

If your binary is missing, your node halts and you start missing blocks. If 1/3+ of voting power misses the upgrade, the chain itself halts.

Manual upgrade (without Cosmovisor)

If you do not run Cosmovisor:

1. Watch the chain. When it logs UPGRADE "v0.3.0" NEEDED at height: X, the next block will halt.
2. Stop the validator: sudo systemctl stop safrochaind.
3. Replace the binary: cp ./safrochaind /usr/local/bin/safrochaind.
4. Start the validator: sudo systemctl start safrochaind.

Done correctly, this loses 1–2 blocks. Done in panic at 03:00 UTC, this loses your validator.

State pruning

Disk grows fast on a synced node. Configure pruning in ~/.safrochain/config/app.toml:

pruning = "custom"
pruning-keep-recent = "100"
pruning-interval = "10"
min-retain-blocks = 100

Mode	What it keeps	Use case
default	last 362 880 + every 100	regular full node
nothing	only the last block	archive node: no
everything	every block forever	archive node: yes
custom	as configured	validator (recommended)

Validators rarely need to query history; aggressive pruning saves disk and IOPS. Pair it with:

[state-sync]
snapshot-interval = 1000
snapshot-keep-recent = 2

So you (or others) can fast-sync new sentries from your snapshots.

Routine maintenance schedule

Cadence	Task
Daily	review Grafana, confirm voting power, check Discord/Telegram alerts inbox
Weekly	withdraw commission, check disk free, review missed-block heatmap
Monthly	rotate logs, patch OS (unattended-upgrades), audit firewall rules
Quarterly	run a paging-test alert, restore-from-backup drill, key-rotation review
On governance proposal	read it, vote: safrochaind tx gov vote <id> yes …

Voting on governance

Validators are expected to vote on every governance proposal: your delegators delegated to your judgement as well as your uptime.

safrochaind query gov proposals --status voting_period --node $NODE

safrochaind tx gov vote <PROPOSAL_ID> yes \
  --from validator --keyring-backend file \
  --chain-id $CHAIN_ID --node $NODE $FEES --yes

Vote options: yes · no · no_with_veto · abstain.

Quick reference card

Action	Command
Show me my validator	safrochaind query staking validator $VALADDR
Am I jailed?	safrochaind query staking validator $VALADDR -o json | jq .jailed
Pending rewards	safrochaind query distribution rewards <addr> $VALADDR
Withdraw commission	safrochaind tx distribution withdraw-rewards $VALADDR --commission …
Edit commission	safrochaind tx staking edit-validator --commission-rate 0.06 …
Edit moniker	safrochaind tx staking edit-validator --moniker "new-name" …
Self-delegate	safrochaind tx staking delegate $VALADDR <amount>usaf …
Unbond	safrochaind tx staking unbond $VALADDR <amount>usaf …
Unjail	safrochaind tx slashing unjail …
Vote	safrochaind tx gov vote <id> yes …