Runbook: Restart & Rebuild

What needs a restart, and what does not.

Nothing you author needs one

You changed	Do
A node, component, app, skill or block, pushed to your org’s Git repo	Nothing. It is live when the universe’s sync applies the commit
Something installed from the marketplace	Nothing. It is live when the install finishes
The platform images	Update them, below

Update the platform images

# On a server
unoverse deploy

# On this machine
unoverse start --pull

Both pull the latest images and restart the services.

Restart a service

docker compose restart unoverse

The workflow engine runs inside the unoverse container, so that one restart covers it.

Full teardown and start

When things are truly stuck:

unoverse stop
unoverse start

Verify

unoverse check

Services, schema and environment. To count the nodes the platform loaded, from inside the container, since :4106 is Docker-internal:

docker compose exec -T unoverse node -e \
  "fetch('http://127.0.0.1:4106/nodes').then(r=>r.json()).then(d=>console.log((d.nodes||[]).length)).catch(()=>console.log(0))"

Troubleshooting

Issue	Cause	Fix
A pushed node or component is not in canvas	The sync has not applied the commit, or hit a lint error	Check the org’s Source tab History, then press Sync now
A component renders its old version	The browser cached it	Hard-refresh
`nodes: 0`	unoverse did not load its packages	`unoverse logs unoverse`

core.md: Initial deployment
test.md: Full health check

Deploy

Operate

Runbook: Restart & Rebuild

Nothing you author needs one

Update the platform images

Restart a service

Full teardown and start

Verify

Troubleshooting

​Nothing you author needs one

​Update the platform images

​Restart a service

​Full teardown and start

​Verify

​Troubleshooting

​Related

Nothing you author needs one

Update the platform images

Restart a service

Full teardown and start

Verify

Troubleshooting

Related