#CLI commands to understand baseline information before upgrade

1 messages · Page 1 of 1 (latest)

fallen creek
#

What CLI commands are available for capturing a full system snapshot before an upgrade? We're aware of status, gateway status, plugins list, cron list. Are there others — like doctor, version info, model config, channel state — that would give a more complete picture? ? I am running 2026.4.20

reef valleyBOT
#

To help others find answers, you can mark your question as solved via Right click solution message -> Apps -> ✅ Mark Solution

weary gulchBOT
fallen creek
#

hello krill

outer templeBOT
#

Let me verify the exact commands from the current docs/source so I don't make any up.

#

I found a newer CLI tree in the current docs too, and there’s actually a backup command family that’s worth including for pre-upgrade snapshots. Let me verify that one before I answer.

#

hey, yep, and the big extra one is:

openclaw backup create --verify

That gives you an actual restorable pre-upgrade snapshot, not just diagnostics.

For a solid baseline on 2026.4.20, I’d capture:

openclaw --version
openclaw config file
openclaw config validate

openclaw status --all
#
openclaw gateway status --json
openclaw health --json

openclaw models status --json
openclaw channels list --json
openclaw channels status --probe --json

openclaw plugins list --json
openclaw cron list
openclaw agents list --bindings
openclaw sessions --all-agents --json
openclaw devices list --json   # if you use paired devices/nodes

A few notes:

#
  • openclaw status --all is the main pasteable support snapshot.
  • openclaw channels list --json gives configured channel/auth state.
  • openclaw channels status --probe --json gives live channel health.
  • openclaw models status --json gives resolved model + auth overview.
  • openclaw doctor is good too, but I’d treat it as a sanity check, not the core snapshot. Plain doctor is fine, just avoid --fix/--repair if you want read-only behavior.
  • If you want usage/quota windows too: openclaw status --usage

If you have important cron jobs, also do:

openclaw cron show <job-id>

for each one, since cron list is summary-level.

Docs:

fallen creek
#

yes

outer templeBOT
#

yep, here’s a one-shot collector:

TS="$(date -u +%Y%m%dT%H%M%SZ)"
OUT="openclaw-snapshot-$TS"
mkdir -p "$OUT"

run() {
  local file="$1"
  shift
  echo "+ $*" | tee -a "$OUT/commands.log"
  "$@" >"$OUT/$file" 2>&1
  printf '%s\t%s\n' "$?" "$*" >>"$OUT/exit-codes.tsv"
}

run version.txt openclaw --version
#
run config-file.txt openclaw config file
run config-validate.json openclaw config validate --json

run backup.txt openclaw backup create --verify

run status-all.txt openclaw status --all
run status-usage.txt openclaw status --usage
run gateway-status.json openclaw gateway status --json
run health.json openclaw health --json

run models-status.json openclaw models status --json
run channels-list.json openclaw channels list --json
run channels-status.json openclaw channels status --probe --json

run plugins-list.json openclaw plugins list --json
#
run cron-list.txt openclaw cron list
run agents-bindings.txt openclaw agents list --bindings
run sessions.json openclaw sessions --all-agents --json
run devices.json openclaw devices list --json

run doctor.txt openclaw doctor --non-interactive
run security-audit.json openclaw security audit --json

tar -czf "$OUT.tar.gz" "$OUT"
echo
echo "Wrote:"
echo "  $OUT/"
echo "  $OUT.tar.gz"
echo
echo "Optional: run 'openclaw cron show <job-id>' for each important cron job."