#OpenCLaw doesn't start anymore

1 messages · Page 1 of 1 (latest)

lost knoll
#

After a restart openclaw can start up:
21:38:28 [INFO] Appending plugin "telegram" configuration
21:38:28 [INFO] Plugin "whatsapp" does not meet requirements, skipping
21:38:28 [INFO] Starting OpenClaw gateway...
21:39:21 [WARN] OpenClaw exited with code 1
21:39:23 [WARN] OpenClaw process exited before device approval could complete
🦞 OpenClaw 2026.4.24 (cbcfdf6) — Powered by open source, sustained by spite and good documentation.

Gateway not reachable. Is it running and accessible?
Gateway target: ws://127.0.0.1:18789
Source: local loopback
Config: /data/.openclaw/openclaw.json
Bind: loopback

olive iglooBOT
#

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

frosty pineBOT
#

That log is the supervisor saying “gateway died”; it doesn’t show the actual crash yet.

You’re on 2026.4.24, and 2026.4.25 is already out with plugin startup/install + bundled runtime-deps fixes, so I’d update first before touching config. (github.com)

Try:

which openclaw
openclaw --version
openclaw update
openclaw doctor
openclaw gateway restart
openclaw gateway status --deep

If this is Railway/container — /data/.openclaw strongly suggests that — don’t delete the /data volume; that’s where state/config live. Redeploy/update the image instead, keeping the volume attached. (docs.openclaw.ai)

#

If it still fails, paste the 50–100 log lines before OpenClaw exited with code 1, especially anything with:

Error
ENOTEMPTY
Cannot find package
plugin-runtime-deps
EADDRINUSE

Emergency rollback for a normal npm install:

npm i -g openclaw@2026.4.23
openclaw doctor
openclaw gateway restart
lost knoll
#

openclaw update
Updating OpenClaw...

│
â—‡ âś“ Updating via package manager (28.82s)
│
â—‡ âś“ Running doctor checks (58.18s)

Update Result: OK
Root: /data/.npm-global/lib/node_modules/openclaw
Before: 2026.4.24
After: 2026.4.25

Total time: 87.36s

Updating plugins...
No plugin updates needed.
Completion cache update failed: Error: spawnSync /usr/local/bin/node ETIMEDOUT

Restarting service...
Gateway aborted: config is invalid.
plugins.entries.feishu: plugin feishu: plugin requires OpenClaw >=2026.4.25, but this host is 2026.4.24; skipping load
plugins.entries.whatsapp: plugin whatsapp: plugin requires OpenClaw >=2026.4.25, but this host is 2026.4.24; skipping load
Fix the config and retry, or run "openclaw doctor" to repair.

frosty pineBOT
#

That’s a split-brain install: the CLI update succeeded to 2026.4.25, but the thing restarting the gateway is still running 2026.4.24.

The important clue is:

After: 2026.4.25
...
plugin requires OpenClaw >=2026.4.25, but this host is 2026.4.24

So don’t edit the Feishu/WhatsApp config yet — fix the binary/service mismatch first.

Please run/paste:

which -a openclaw
# 
openclaw --version
openclaw gateway status --deep
openclaw config get meta.lastTouchedVersion

Then try the repair path from the docs:

openclaw doctor --fix
openclaw gateway install --force
openclaw gateway restart
openclaw gateway status --deep

If this is Docker/Railway/container-based, openclaw update inside the container may not be enough — you likely need to redeploy/rebuild the container image so the service itself starts 2026.4.25, while keeping /data attached.

#

Also:

Completion cache update failed: spawnSync ... ETIMEDOUT

is probably not the gateway crash; it’s just shell completion cache generation timing out. The real blocker is the service still launching 2026.4.24.

lost knoll
#

This Openclaw is running on a VPS in a docker container.

frosty pineBOT
#

Yep — then don’t use openclaw update inside the running container. Docker containers are image-based, so you now have mixed state:

  • CLI/global install in /data/.npm-global = 2026.4.25
  • gateway process baked into container image = 2026.4.24

Fix is to update/recreate the Docker image/container, keeping /data mounted.

On the VPS host, in the compose directory:

docker compose logs --tail=200 openclaw-gateway

Then either:

If you build from repo:

# 
git pull
docker compose build
docker compose up -d --force-recreate openclaw-gateway

If you use the prebuilt image:

docker compose pull
docker compose up -d --force-recreate openclaw-gateway

If your compose file pins ghcr.io/openclaw/openclaw:2026.4.24, change it to 2026.4.25 or latest first.

After recreate:

# 
docker compose logs -f openclaw-gateway

Important: do not delete /data / the mounted volume — that’s your config/auth/session state. The container itself is disposable; the volume is not.

lost knoll
#

docker compose pull worked
docker compose up -d --force-recreate also worked
openclaw-gateway didn't work on the VPS:
no such service: openclaw-gateway

frosty pineBOT
#

No worries — that just means your compose file uses a different service name than the docs example. My openclaw-gateway name was an assumption.

Run this in the VPS compose directory:

docker compose config --services
docker compose ps

Then use the actual service name it prints. For example, if it prints openclaw, run:

docker compose logs --tail=200 openclaw
docker compose logs -f openclaw

Or just show all logs without naming the service:

# 
docker compose logs --tail=200

To verify the container image/version:

docker ps --format 'table {{.Names}}\t{{.Image}}\t{{.Status}}\t{{.Ports}}'

If it’s now running 2026.4.25, the earlier plugin version error should be gone. If the gateway still exits, paste the new docker compose logs --tail=200 output — that’ll show the real remaining crash. Docs: https://docs.openclaw.ai/install/docker

lost knoll
#

Works again!

frosty pineBOT
#

Nice, glad it’s back! 🎉

For the thread record: the fix was updating/recreating the Docker image/container, not openclaw update inside the container — that had caused a 2026.4.25 config/plugin state with a 2026.4.24 gateway binary.

If this fixed it, please right-click the solution message → Apps → ✅ Mark Solution so Answer Overflow indexes it for the next person.