wip up

wip up

Brings the stack up: builds what needs building, creates the network, starts sidecars, mirrors the source, and starts the primary container.

wip up [-d] [--no-sync] [--no-cache] [--watch] [--interval N]

What it does, in order

Under mode: container and mode: compose-native:

1. Validate --interval (only when --watch is given) — before any side effect, so a bad value doesn't leave a half-started stack behind.
2. Build compose-native build: services — once per invocation, in depends_on order. No-op in container mode. See Compose Build.
3. Create the network if network: is set and missing. See Networking.
4. Start every sidecar, by name. Existing ones are started, missing ones created.
5. Mirror the source into the sync volume, if sync: is configured and --no-sync wasn't passed. See Source Sync.
6. Start the primary container — start if it exists, run --name … if not.
7. Enter the watch loop, if --watch was given.

$ wip up -d
wip: creating network 'app-tier'
wip: dependency 'redis' not found, creating it
wip: starting existing dependency 'development.mysql'
wip: syncing /home/me/app -> app-src:/app
wip: run `wip sync --watch` in another terminal to keep /app up to date
wip: container 'app' not found, creating it

Under mode: compose it's much shorter: mirror the source (unless --no-sync), then delegate to <compose command> -f FILE [-p PROJECT] up [-d].

Flags

-d / --detach

Run the primary container in the background. Without it, wip attaches to the container and your terminal is held until it exits.

Sidecars are always started detached, regardless of this flag.

--no-sync

Skip step 5. Useful when you know the volume is already current and want a faster boot, or when you're about to run wip sync --watch anyway.

Has no effect without a sync: block.

--no-cache

Passes --no-cache to the compose-native image builds in step 2. No effect in container mode (nothing is built there) or compose mode.

--watch / -w and --interval N

Poll every N seconds (default 5) and restart any dependency that has exited and whose restart: allows it:

$ wip up --watch
wip: watching app, mysql for exited restart: containers every 5s (running detached; Ctrl-C to stop)
wip: 'mysql' has exited, restarting it (restart: unless-stopped)

Key points:

• Implies -d. The primary container can't hold an attached TTY while the loop polls on the same thread.
• Foreground loop. Ctrl-C stops supervision; closing the terminal does too.
• Not available under mode: compose — there's no service list to poll.
• Status-based, not event-based — it can't distinguish a crash from a wip stop you ran elsewhere.

Full semantics and limitations: Restart Policies.

--interval must be positive:

--interval must be a positive number

Idempotence

Re-running wip up against a running stack is safe. Existing containers are started, which is a no-op when they're already running. Nothing is recreated.

That also means config changes don't apply to existing containers. New ports, volumes, or env vars require:

wip down && wip up -d

Common problems

Symptom	Likely cause
App can't resolve redis / db by name	network: unset — see Networking
New ports: entry isn't listening	container predates the change; wip down && wip up -d
container: must be set…	Dependencies
pull access denied	Registry Authentication
no matching manifest for linux/…	Architecture Mismatch
Boot hangs with low CPU	bind-mount overhead — Fixing a Slow Boot
0x8007000e / too many mounted volumes	Volume Limit Reached

Related

• wip stop / wip down
• Restart Policies
• Source Sync
• Debug Output — wip up --debug when boot is slow