If @bobbyD and the team is investigating, and trying to figure out a root cause of your "odd" and "interesting" troubles, they may not know the answer within a week. Even if they do know the root cause, it may take more than a week to code a fix. Maybe it even requires a fix outside of the Hubitat firmware itself?
2 Likes
Honestly, I don't care if it takes a week or a month or more to pin it down ... but considering that there has been zero communication since @bobbyD said he was pulling in someone else, despite multiple followup inquiries from me, right now, it feels like the issue was dropped. if that's the case, maybe it deserved to be .. but a note that my case was being closed seems like a reasonable expectation in that event. Just say something. Anything.
Complete and utter silence isn't useful and I have no idea if the problem has been given 3 seconds of thought during the last 9 days or become someones sole focus. I've been patient and optimistic so far; but not because anyone has reached out and given me any reason to be.
At this point, I still have no idea if there is even any interest in figuring it out let alone if someone is working on it.
We have tried to monitor your hub but as mentioned before you'll need to continue to use the C-8, if you'd like us to help you troubleshoot. There was nothing obvious in your hub's engineering log, which may point to a mesh problem. Usually after migration it may take 3 to 5 days for the mesh to settle. In your case, with as many devices you have, it may take even longer. I suggest moving back to the C-8 and give it some time to see if the problems persist.
1 Like
Funny you should mention that .. On the PM thread I specifically asked if you still needed access to the C8 - 7 days ago or I was going to need to roll back and got no response to that either .. I finally gave up waiting and rolled back to the C5 4-5 days ago as I need things to work, or I need to know something is happening in the background while I'm inconveniencing my household.
I've offered to coordinate diagnostics, perform tests, anything HE wants me to do ... but in the last 9 days this is the first acknowledgement of anything.
So, lets discuss it here instead so others can weigh in.
Could it be mesh? Sure, however ... Every time I move the mesh to the C5, it takes about 5-10 minutes, everything works, and everything is just as reliable and solid as it's always been. I switch the C8 back in, and the entire system goes down the toilet.
We're not talking about slowness, an occasional miss, or generally frustrating behavior ... the mesh is dead as a door nail from the hub's perspective, and nothing can be commanded for hours at a time or more. I did deal with it for nearly a 5 day stretch at one point (much to my family's chagrin), there was no improvement seen - in fact, I'd venture to say it seemed worse.
The fact that the same mesh, unchanged, works fine on the C5 and not the C8 leans me away from thinking the mesh itself is the problem (maybe in error). I do, however, accept that there may be a difference in the way the 500 series chip vs the 800 series chip interacts with it. In that case though, I don't think any amount of "settling" is going to make a difference, do you? (I also mentioned this in the same PM from 7 days ago where I also mentioned several other possibilities, and suggested two possible approaches to testing)
I'm willing to do pretty much anything you want - besides plug it back in and cross my fingers. As you pointed out, there are a lot of devices connected to my mesh ... and they all have things to do; many of them are only controllable from the hub - so simply leaving the C8 connected in perpetuity is not really an option.
I think we're beyond a "wait and see" approach. Been there, done that ... the problem I see with this theory is that even if it can take days for the routing to optimize, "dead" shouldn't be the result of insufficient routing, and if it was, I'd expect the reverse to be true when reverting to the C5 after days on the C8. That's not the behavior I see.
So besides plug it back in ... do you have any more ideas? The only thought I have at the moment is doing another backup from the C5 and a fresh restore to C8 after a factory reset. However, if you or the community think other tests or diagnostics done on the C8 in its current state could be beneficial, I will leave it intact for the moment.
I've been thinking about suggesting that for a few days but didn't want to suggest something that flushes any clues already acquired down the drain.
I had once chance to do a full, automated Migration... from a C-8 to a C-8. Worked perfectly, of course. Be aware that a restore does a soft reset first, which is one of the first steps of 'factory reset'. Restoring a cloud backup, if it contains the radio data, offers to restore them or not. If you do, then it automatically resets the radios SOC first. Which is a second step of 'factory reset'.
I'm in the "try another cloud migration" camp. As long as you're not excluding / including devices it seems like picking which Hub is powered up is going to give you a quick backout to the C-5 if another migration attempt isn't beneficial.
1 Like
Not necessarily. Where is the C8 hub located in relation to the C5? Are they in very close proximity? Based on the details you provided before, and in absence of any indication in your engineering logs that the radio has a problem, I am inclined to believe that you are dealing with a bandwidth issue. I'd check the Logs for hyperactive devices. Here is a post that might give you some ideas what to look for:
When the C8 and the C5 swap active duty, they sit in the exact same 3”x3” spot. Even before I had issues, I was only introducing a single variable change intentionally.
I think you’ll need to elaborate on the possibility of a bandwidth issue. If anything, I’d assume better bandwidth since more devices would have single hop access to the hubs dedicated antennae, and fewer packets needing to bounce around the mesh.
Though I would still like to understand how you think bandwidth could be contributing here, I'm going to move forward since there doesn't seem to be anymore feedback coming. I will probably reset the C8 in the next day or two and repeat the migration process.
I've since reconnected my zigbee devices to the C5 and the zwave mesh has been stable and reliable for over a week. I will run a fresh cloud back up, and then perform the restore to the C8.
@bobbyD ... Do the engineering logs survive reboots? If so, can you please pull them and see if there are any new clues?
I reset the C8 to factory and restored the a new backup several days ago. The scenario is now different, but not much better. Zwave seems to slow down and become less reliable as the day progresses, and overnight it completely fails requiring the hub to be rebooted in the morning. I don't have to pull the power, just reboot. It hasn't been sporadically failing and recovering like before.
60+ overall... (except for 2 of my iblinds which I didn't notice the batteries died... so need charged 
. A lot of people have way more than me
Yes
That isn't a common symptom for Z-Wave network, but rather hub overload. Are Zigbee devices not suffering from same problem, or any other connected devices for that matter?
Actually, when I did the restore this time, the Zigbee devices don't even work ... so I don't have them as a point of comparison now.
There isn't much in the way of other connected devices that I can test ... but the GUI itself doesn't render any more slowly than it usually does (it's never quick), so on the surface things seem ok. I used to have a URL that showed the CPU utilization of the hub ... not super useful for most diagnostics, but did occasionally add a tidbit of useful info during troubleshooting - I can't remember what it was though. Perhaps that would expose a trend of increasing utilization throughout the day if I could find it?
Thanks ... I'm chasing a theory that may go nowhere, but some of your other comments on the forum about your environment are similar to mine ... so I figured what you have running stable might be a useful reference.
Absolutely! I will always stand up for his unmatched instability and uselessness! A role model for us all... 
2 Likes
Is there a way to actually disable the WIFI module? I mean legitimately turn it off, not just have no configuration on it.
It turns off when hard wired I believe.
It may not be transmitting anything, but I highly doubt the module is actually powered down. I suspect it’s vampiric and constantly pulling some amount of power from the system (even if it’s a small amount).
I’m leaning towards a power issue right now, but don’t have enough data points to prove it … so I’m trying to figure out if I can save any someplace else and have a little more to go around for what’s actually using it.
Typical WiFi modules, when not transmitting, draw about 100mW of power in Receive mode. Since this module is not even connected to a WiFi network, I doubt it is even using more than a few mW of power.
What is the concern? The whole hub uses less that 5W of power at peak usage. I am just genuinely curious why you're concerned about the vampiric draw of an unused WiFi module? Are you running 'off the grid' using solar/wind/batteries for your home?
2 Likes
