Lorawan-stack: Issue with redis, all uplinks are dropped

[alexbn71]

Summary

The newest versions have an issue related with redis that drops all uplinks messages received by nodes.

Steps to Reproduce

I use docker-compose.yml with this settings for redis:

  redis:
    image: redis
    command: redis-server --appendonly yes
    restart: unless-stopped
    volumes:
      - ${DEV_DATA_DIR:-.env/data}/redis:/data
    ports:
      - "127.0.0.1:6379:6379"

...
...
...

   environment:
      - TTN_LW_REDIS_ADDRESS=redis:6379

What do you see now?

dropped uplinks in console and this message in the stack:

WARN Finished unary call duration=14.29155ms error=error:pkg/redis:store (store error) error_cause=dial tcp 127.0.0.1:6379: connect: connection refused error_correlation_id=91199bf942f5474f947218d76407fab5 error_name=store error_namespace=pkg/redis grpc_code=Unknown grpc_method=HandleUplink grpc_service=ttn.lorawan.v3.GsNs namespace=grpc request_id=01E46M0W9GGPMWS1FE0G6JHP55

What do you want to see instead?

that it works like in v3.6.0

Environment

the issue appear after an upgrade to v3.6.1 or v3.6.2.

[KrishnaIyer]

@alexbn71 : Is this happening for v3.6.2 as well? What steps did you take to update?

[alexbn71]

with a _docker-compose pull_ after this change in my docker-compose.yml:

  stack:
    image: thethingsnetwork/lorawan-stack:3.6.2

and the console confirm that I'm on v3.6.2...

[KrishnaIyer]

Ok I see the issue. Will be fixed in v3.6.3. Also you can update to our pre-release new minor version v3.7.0-rc2 where this is resolved.

[alexbn71]

Thank you @KrishnaIyer

[alexbn71]

does this update require a db migration?

[KrishnaIyer]

If you update to v3.7.0-rc2 yes. It does. You can also wait for v3.6.3 which doesn't need one.

[alexbn71]

and if I migrate the db can I still back to v3.6.0 in case the issue is still there?

[KrishnaIyer]

Yes. Downgrading won't be a problem.

[alexbn71]

Upgraded to v3.7.0-rc2 and I can confirm that it works now! Thanks.