Hub Mesh issue following 2.2.6 upgrade

🛎️ Get Help
1

Currently have two C7s. Upgraded both to 2.2.6.133 last night (first upgrade from 2.2.5.x for both).

Hub A, the main hub, has all of my devices (vast majority zwave) except for one. Hub B has one device, a S0 zwave lock (only S0 device I have). Hub B shares the lock with Hub A via Hub Mesh. I've had no issues with this since setting up Hub Mesh.

Following the 2.2.6 upgrade on both, Hub B's Hub Mesh setting page started showing the 500 error, and Hub A is no longer able to control the shared device. Through multiple reboots of both devices, I notice that the 500 error starts showing on Hub B's Hub Mesh page as soon as that hub is viewable within Hub A's Hub Mesh settings page.

I've tried creating new tokens (for both hubs) and I've removed / recreated the share on the lock multiple times. The same cycle occurs regardless of a shared device or not: as soon as Hub A can see Hub B within it's Hub Mesh settings page, Hub B starts showing the 500 error on it's Hub Mesh page, and if the device is shared, it cannot be added back to Hub A (though is viewable, but not selectable, in the link menu).

I tried reverting Hub B back to 2.2.5.x, same issue (though I believe both hubs need to be running the same version for Hub Mesh to work... I have not reverted Hub A back to 2.2.5.x). What else am I missing?

2

Are there any logs when the problems occur?

3

No logging, no system events.

4

There was an issue reported in November that seems to have similar flavours... It was fixed by deleting a rule which I find odd...

Tagging @gopher.ny in case the current issue is familiar...

5

I see hub with id starting with 3e complaining about space in the hub token, but that was yesterday.
If the issue still exists, could you please click Refresh several times so that the 500 error generates several identical/similar entries in the log file?

1 Like
6

== Hub B

Done. Other info that may be helpful:

  1. Both hubs are now running 2.2.6.134.
  2. I'm now realizing that I'm using TCP when things are getting to where they are currently. When using UDP, the hubs are never able to see one another. I cannot recall if I have always had them using TCP or not... nothing has changed with my network topology though, and nothing new has been added or removed since this issue has begun.

Appreciate your attention!