Scene says it ran, three of five devices not responding

I have a scene activated by SAR that opens five blinds (iBlinds) five minutes after sunrise. Usually runs w/out a hitch. Got up this AM and three of the five blinds hadn't opened. Hub is on 2.2.6.130.

The three blinds that didn't open show no events when the scene was activated at 6:47 (sunrise 6:42 + 5 minutes). For example, event log below for one of the three blinds that didn't open below - all three blinds that didn't open show no events at or around the opening time:

Logs show that SAR ran (Blinds all open at sunrise +5), the scene was turned on (Blinds alll open scene was turned on), and setLevel commands were sent for all five blinds, but only two seemed to get the commands:
2021-03-28 08_22_36-Logs

Z-Wave details is clean, no ghosts, no changes to Z-Wave devices, all are connected to power and reporting normal routing, four of the five connect direct to the hub, one routes through one node, all connect at 40k.


I have a Pico set up that opens the three office blinds using setLevel, and used that when I got up this AM to open the blinds and that worked normally. I used Google voice command to open the other blind that hadn't opened.

I know that HE was looking at issues where multiple commands sent by a scene were getting lost. Wondering if this is an example of that an any info I can provide to help trouleshoot, @bravenel.

This isn't surprising if it didn't work. Events, generally speaking, don't show when commands were sent, only when a report is received from the device they something actually changed. (Technically, that's up to the driver, but pretty much all Z-Wave and Zigbee drivers work like this; iBlinds' custom driver, which I don't like for a variety of reasons, is an exception that "optimistically" generates events.) If you're using my driver, it will print a debug log when a command is sent if those are enabled. Lots of drivers don't do that either, though. However, the Scene app logs suggest they were sent.

So that still leaves the question of why: my guess is just a temporary problem with the device. As you know, I habe a lot of these, both v2 and v3, and have used them with both a C-7 and earlier hub model. Every once in a while, I have one that just doesn't respond. Trying again even a few minutes later usually works. Luckily, all of mine seem to be fine lately, but when they don't work, I've always had a hard time figuring out why (mesh issues? device/firmware/wakeup bug?).

1 Like

Looking at many posts.

A few people seem to have good luck with devices always doing what's expected. Others, including me, don't.

Even after looking at all the normal problem causes (ghosts, etc.), there isn't anything obvious.

There's typically the catch-all of "mesh problems", but that doesn't lead to a solution. With z-wave plus devices, the mesh is always adapting (unlike the original version, where it took a network repair to reset the paths).

My issues surface in long delays and non-responsiveness. At times, it's almost like the hub's radio isn't listening or transmitting for periods. Using a protocol monitor, I tend to see routing errors during those times.

Given the nature of zwp, problems can explode exponentially. If a packet doesn't get ACK'd, the device will resend it and start a semi panicky attempt to reach the destination by chatting up all it's neighbors. Once that happens to a few devices, you can imagine the traffic. That's kinda my gut feel about things--something goes south, then things explode until everything settles down again.

Thanks for the details. I'm using your driver (v3) and I'll go make sure the logging is turned on! We'll see what happens tomorrow morning...

And of course everything ran perfectly this morning. :slight_smile:

I have turned on debug logging in your driver, will leave it on for a few days to see if I can "catch" another event like yesterday.

1 Like

Note that, like most Hubitat drivers, it should auto-disable after 30 minutes. There is a line in the code you can modify to increase this time or eliminate this altogether, though I don't have it in front of me at the moment to say exactly what. (But while it's probably good to check anyway, my guess is it won't be surprising: driver probably getting command, devices probably not responding. :slight_smile:)

1 Like

Thanks - I'll see if I can hunt down the line and modify it to last 8 hours, and set it again before I go to sleep. That should cover it. Though I'm doubtful it will fail again now that I'm watching... :wink:

2 Likes

It will not fail until exactly 30 seconds after debug logging turns off :wink:

2 Likes

LOL..perfect truth. :rofl:

1 Like

Damn scene is running perfectly again. :wink:

So it appears it was an inexplicable glitch that I'll write off to sunspots and Wiccan meddling in/my mesh by jealous neighbors who wish they had iBlinds. :smiley:

1 Like