FW Update for Zwave stability (Ezlo Linux firmware [production-155])

Hi everyone,

I wanted to give a bit more information about the zwave stability improvements we have done on the latest FW update.

  1. Control of Set command

When you had meshbots with several zwave device actions, some actions did not get executed properly.
We were just sending the set command and forget about it being executed on device. We were relying on zwave libraries to handle that. Unfortunately it created problems since there may be lost or jammed signals.

We repeat the command when necessary. we manually check if ACK response comes back (confirm message is received by device) If the ACK is not received, we retry it 3 times if still not received then this node will be marked as unreachable. Then it goes to a suspended state of 1 hour. (Messages won’t be send to same node until it becomes reachable again)

  1. Meshbot Execution

When you use the default execution plan in Meshbot actions (sequential) , a set command which was being tried to send to an unreachable device took almost 30 sec or more since it was trying all possible mesh routes to reach the device. And it was blocking execution.

We added prevention of sending message to unreachable nodes too often.
If it is unreachable we block sending messages to this device from everywhere for x hour (1 hour). It will still be tried every 1 hours until it becomes reachable. And this retry mechanism won’t block meshbot execution.

  1. Periodic Reconstruct

Everyday at 3am. we are performing an automatic reconstruct of zwave network. Reconstruct means finding out shortest ways to each node on the network considering unreachable nodes. So it will allow faster execution of set commands to devices.

So if there are still customers who did not upgraded to latest version ( then we strongly recommend upgrading and then let us know if you see improvements or let us know any feedback on the topic.

Best regards.


When exactly was the firmware update? I ask because my controller restarts once a day as a workaround for a bug that still needs to be fixed (Request #299707). That’s why I can’t estimate this based on the controller runtime.

It was on September.
You can see in release notes:

This is very good news and a welcome update. I was suspicious the canned ZWave libraries wouldn’t handle some of the subtle error legs needed. Love it or hate it Vera was quite good at ZWave, now Ezlo should be in better alignment to what we had with Vera.

Do you show any visible indication of a node being in a suspended state? I believe Vera had a red dot on the GUI for this. Further are you saying after an hour you reset the suspension and presume the node is now ok?

I assume it says “Unreachable” on that particular device in the Ezlogic web UI, Settings Devices area as the indicator.

This new update sounds like good news also.

The node will be shown as unreachable. But the meshbot execution will retry after 1 hour.
If meanwhile the node comes back itself then it will become reachable again. So there is no suspension state actually. Just retry interval will be longer.