Z-wave wake-up detection doesn't seem to make sense...

[TC1]

Hi Steve,

I've had this weird issue for a while with the plug-in but have been ignoring it, decided to collect some data today. I have a z-wave multi-sensor that wakes approximately every 12.5 hours, normally the PI picks that up and records it appropriately. But it seems if I restart HS4 (and hence the PI) it throws everything off and I start getting wake-up missed alerts.

Below is the status of the SDJ-Monitoring device, if you notice the PI recorded the last wake at 1/8 3:51PM. Later that afternoon I did reboot HS4.


But if you look at the log below, HS4 did indeed process a wake-up (z-wave node 22 is the device) about 12.5 hours letter as usual. But then at 6:34am the PI claims the device missed the wake-up????


Here are the reference IDs for the actual monitored and monitoring devices:



It doesn't seem to make sense? Is the PI losing information during reboots? I assumed everything was written to a database to keep track. Why didn't the PI record the wake-up at 4:30am?

[SteveMSJ]

Hi Steve,

I've had this weird issue for a while with the plug-in but have been ignoring it, decided to collect some data today. I have a z-wave multi-sensor that wakes approximately every 12.5 hours, normally the PI picks that up and records it appropriately. But it seems if I restart HS4 (and hence the PI) it throws everything off and I start getting wake-up missed alerts.

It doesn't seem to make sense? Is the PI losing information during reboots? I assumed everything was written to a database to keep track. Why didn't the PI record the wake-up at 4:30am?

Hi,

You are correct, it doesn't seem to make sense The information is stored in the PED of the SDJ-Health root/parent device, so unless HS is failing to save its database of devices correctly on a reboot, no information should be lost.

Can you un-filter your log around the 4:30AM time point a look at all the messages around that wake-up.

The first message from the Z-Wave-plug-in should be '...Wake-Up notification received for Node 22', which should be just before the 'processing' message you show in your screenshot. This is the message that the plug-in looks for.
There should then be a message from SDJ-Health saying 'Wake-Up Device #62 added to the message queue for processing'. However, you will only see that if LogLevel is set to at least 1. It looks like yours is 1 because I can see messages further up your log.

The SDJ-Health message appears to be missing from your screenshot. I don't know why that should be if SDJ-Health was running at the time. However, if the 'received' message is missing that would explain it, but that seems unlikely.

Let me know what you see in the log.

Steve

[TC1]

Sure, here's an unfiltered view before and after the the Z-wave wake-up:

[SteveMSJ]

Sure, here's an unfiltered view before and after the the Z-wave wake-up:

Well that explains why there is an issue, the Z-Wave plug-in is failing to log the '...Wake-up Notification received...' message on your system. I have been doing this a long time and I have never seen that. The only similar issue I have seen is when the Z-Wave plug-in gets in an unstable state, you start getting high cpu and delays and it stops logging wake-ups all together.
From what you describe your issue only happens after a reboot and then after a while the Z-Wave plug-in settles down. Out of interest can you search back through your log to the wake-ups before the reboot and check they are showing the 'received' messages. Similarly further forward in the log do they start showing again? Let me know what you find.

Whilst I've never seen this, or had it reported before, it sounds like it happens consistently every time you reboot on your system. I doubt we will get anywhere with HomeSeer asking them to track down the issue with their plug-in so the best thing would be for me to put in a work around in my plug-in. If you can confirm from your log that it is consistently not logging the Wake-up Notification Received messages but still logging the Wake-up Notification Processing messages I can put in a fix.

Steve

[TC1]

...and you nailed it, Steve. Approx 12.5 hours later than this morning the device woke, and this time Z-wave logged a complete "woke" message set. Something is definitely weird. Other than this, my HS4 system operates great, low-cpu usage, no extraneous errors in the logs, etc.

[SteveMSJ]

...and you nailed it, Steve. Approx 12.5 hours later than this morning the device woke, and this time Z-wave logged a complete "woke" message set. Something is definitely weird. Other than this, my HS4 system operates great, low-cpu usage, no extraneous errors in the logs, etc.

Ok. I’ll release a beta with a work around for you to test in a day or so.

Steve

[SteveMSJ]

Ok. I’ll release a beta with a work around for you to test in a day or so.

Steve

TC1 I've uploaded 3.1.0.9 beta with a work around that should hopefully fix your issue. I can't test it as I can't replicate the issue so please could you test it and report back

I sometimes have problems with the updater in HS4 for legacy plug-ins. If you find when you click on the update icon the screen dims but it goes no further, just go to Plug-ins ADD and install the beta as if it was new. It will just update the existing, it won't overwrite anything.

Steve

[SteveMSJ]

TC1 I've uploaded 3.1.0.9 beta with a work around that should hopefully fix your issue. I can't test it as I can't replicate the issue so please could you test it and report back

TC1 Have you had a chance to test? I’d like to release this version from beta if it solves your issue.

Steve

[TC1]

Hey Steve, sorry for the delay, unfortunately once Mondays roll around the "real" job keeps me occupied around the clock (my responsibility is global) and the spouse caught Covid (despite being triple vaxxed) so I'm doing all the cooking and cleaning while she isolates/recuperates. Leaves very little time for the things I really want to do, lol. Now that the weekend is getting here I should be able to start testing this today.

[SteveMSJ]

Hey Steve, sorry for the delay, unfortunately once Mondays roll around the "real" job keeps me occupied around the clock (my responsibility is global) and the spouse caught Covid (despite being triple vaxxed) so I'm doing all the cooking and cleaning while she isolates/recuperates. Leaves very little time for the things I really want to do, lol. Now that the weekend is getting here I should be able to start testing this today.

No problem and commiserations on your household Covid situation.

Steve

[TC1]

Got the latest beta installed, waited for it to stabilize, then rebooted my HS4 system. Unfortunately I just missed the last wake-up from my one z-wave device that does wake-ups, so will have to wait for another 12.5 hours when it should report it with the incomplete wake message in the logs. Should be around midnight my time (UTC -5) so I should have some observations and results for you tomorrow morning my time.

[TC1]

Latest beta seems to have solved issue. Like clockwork, it seems the Z-wave plugin always does not register the first "Wake received" message after a reboot. Here's the sequence of the first wake-up below, with the initialization message unaccounted for:



But this time the plugin did register the wake up and process it:


Nice work!

[SteveMSJ]

Latest beta seems to have solved issue. Like clockwork, it seems the Z-wave plugin always does not register the first "Wake received" message after a reboot. Here's the sequence of the first wake-up below, with the initialization message unaccounted for:



But this time the plugin did register the wake up and process it:


Nice work!

Thanks for getting back to me and I’m glad it sorts the issue.
It is bizarre that the Z-Wave pi does that on your system. I have never seen it and I don’t think anybody else has reported it. Actually I have a vague memory of someone else getting missed wake-ups after a restart but the thread went dead. I will have a look back for that.

I would have suspected a busy network but as it happens to you consistently on only the first wake after a restart I think that can be ruled out. I guess we will never know☹️
Steve

[TC1]

Thanks for getting back to me and I’m glad it sorts the issue.
It is bizarre that the Z-Wave pi does that on your system. I have never seen it and I don’t think anybody else has reported it. Actually I have a vague memory of someone else getting missed wake-ups after a restart but the thread went dead. I will have a look back for that.

I would have suspected a busy network but as it happens to you consistently on only the first wake after a restart I think that can be ruled out. I guess we will never know☹️
Steve

Agreed, this is strange and it bothers me being an engineer, everything should have a cause-and-effect. Being that HST is supposedly working on releasing a new 4.x z-wave PI it might be pointless reporting this phenomenon to them.

[SteveMSJ]

Agreed, this is strange and it bothers me being an engineer, everything should have a cause-and-effect. Being that HST is supposedly working on releasing a new 4.x z-wave PI it might be pointless reporting this phenomenon to them.

Yes, I would not expect HST to devote any effort to it as, there don't appear to be any serious consequences and there are far more important issues for them to solve😐.

I wouldn't expect the HS4 native plug-in very soon, after all it must be a couple of years late already. At the HS4 alpha stage ,conversion of the Z-Wave pi was going to set the example for us developers to follow. To be fair to HS I think changes by Sigma Designs meant that a complete rewrite of the z-wave pi was going to be necessary anyway and therefore there was no point in converting the existing one to HS4. I could be completely wrong on that, but either way when it finally comes, expect a whole new set of issues😁

Steve