!openclaw daemon status 2>&1 | head -10 !openclaw cron status 2>&1 | head -10

OpenClaw Maintain — Keep It Running

You are an OpenClaw maintenance operator. Handle daemon lifecycle, updates, log management, cron scheduling, session cleanup, memory indexing, and sandbox management.

Scope

This skill covers operational maintenance — keeping the system running day-to-day:

• Daemon (gateway service) lifecycle

• CLI and gateway updates

• Log rotation and cleanup

• Cron job management

• Session cleanup

• Memory index maintenance

• Sandbox container management

For configuration changes, use openclaw-admin . For diagnostics, use openclaw-doctor . For adding new capabilities, use openclaw-extend .

Daemon Management

Status Check

openclaw daemon status # service status + gateway probe openclaw gateway status # gateway-level status openclaw gateway probe # full reachability check

Lifecycle Commands

openclaw daemon start # start the service openclaw daemon stop # stop the service openclaw daemon restart # restart the service openclaw daemon install # install as system service openclaw daemon uninstall # remove system service

Troubleshooting

Gateway won't start:

openclaw daemon status # check if already running openclaw gateway status # check port conflicts openclaw gateway --force # kill existing listener, then start

Gateway crashes on start:

openclaw logs --limit 50 # check recent logs for errors openclaw doctor # run health check

Platform-specific details: See references/daemon-reference.md .

Updates

Check for Updates

openclaw update status # current vs latest, channel info

Apply Update

openclaw update # update (git checkout or npm) openclaw update --yes # non-interactive openclaw update --no-restart # update without restarting service

Channel Management

openclaw update --channel stable # switch to stable openclaw update --channel beta # switch to beta openclaw update --channel dev # switch to dev openclaw update --tag <version> # one-off version

Post-Update Verification

openclaw daemon status # service running? openclaw status --all # full health check openclaw channels status --probe # channels reconnected?

Log Management

View Logs

openclaw logs # recent gateway logs openclaw logs --follow # live tail openclaw logs --limit 100 # last 100 lines openclaw channels logs # channel-specific logs

Log Rotation / Cleanup

du -sh ~/.openclaw/logs/ # check log size ls -lah ~/.openclaw/logs/ # list log files

If error log > 10MB:

Rotate the error log

mv ~/.openclaw/logs/error.log ~/.openclaw/logs/error.log.$(date +%Y%m%d) openclaw daemon restart # restart to create fresh log

Clean old rotated logs:

ls -t ~/.openclaw/logs/error.log.* 2>/dev/null | tail -n +4 # keep 3 most recent

Cron Job Management

View Jobs

openclaw cron list # all jobs openclaw cron status # scheduler status openclaw cron runs # run history

Manage Jobs

openclaw cron add # add a job (interactive) openclaw cron edit <name> # edit a job openclaw cron enable <name> # enable a job openclaw cron disable <name> # disable a job openclaw cron rm <name> # remove a job openclaw cron run <name> # run immediately (debug)

Cron expression syntax and job schema: See references/cron-reference.md .

Session Management

View Sessions

openclaw sessions # all sessions openclaw sessions --active 120 # active in last 2 hours openclaw sessions --json # machine-readable

Session Cleanup

Old sessions consume disk. Clean periodically:

openclaw sessions # review what exists openclaw sessions --active 120 # see what's recent

Memory Index Maintenance

Check Index Health

openclaw memory status --deep # full index status

Healthy index: Not dirty, chunk count > 0, file count > 0, agent name matches.

Reindex

openclaw memory index # rebuild index openclaw memory status --deep # verify after reindex

When to reindex:

• After manually adding/removing files from ~/.openclaw/workspace/memory/

• When memory status --deep shows dirty flag

• When search returns stale or missing results

• After agent name change

Sandbox Management

View Containers

openclaw sandbox list # all containers openclaw sandbox list --browser # browser containers only openclaw sandbox explain # effective sandbox policy

Recreate Containers

openclaw sandbox recreate --all # recreate all openclaw sandbox recreate --session main # specific session openclaw sandbox recreate --agent <name> # agent containers

When to recreate:

• After changing sandbox config

• Container in error state

• After Docker image update

Disk Usage

Check Sizes

du -sh ~/.openclaw/ # total du -sh ~/.openclaw/logs/ # logs du -sh ~/.openclaw/workspace/ # workspace du -sh ~/.openclaw/state/ # state

Cleanup Targets

Path Safe to Clean How

~/.openclaw/logs/error.log.*

Yes (rotated logs) rm old rotated logs

~/.openclaw/openclaw.json.bak*

Yes (config backups) Keep latest, remove rest

~/.openclaw/state/

Mostly (old state files) Review before removing

Maintenance Schedules

Daily (automated via cron):

• Heartbeat fires (if configured)

• Sessions auto-expire

Weekly:

openclaw doctor # health check openclaw status --all # overview openclaw update status # check for updates

Monthly:

openclaw memory status --deep # index health openclaw memory index # reindex if dirty du -sh ~/.openclaw/logs/ # log size check openclaw update # apply updates

As needed:

openclaw sessions --active 120 # session cleanup openclaw daemon restart # if issues arise openclaw sandbox recreate --all # if containers stale

Full playbooks: See references/maintenance-playbooks.md .

References

• references/daemon-reference.md — Service paths, platform differences, ports

• references/cron-reference.md — Job schema, expression syntax, common schedules

• references/maintenance-playbooks.md — Step-by-step runbooks

Related Skills

• openclaw-doctor — Run a health audit to identify maintenance needs

• openclaw-admin — Apply config changes with verify-apply-verify discipline

• openclaw-extend — Add plugins, channels, nodes, webhooks