Per the subject and after a bit of research, I believe Hubitat C8 (other vers?) has a bug. Symptoms: after either an extended power outage or a brown out, the C8 refuses to restore itself to normal admin behavior. No response locally or remotely. A number of other posts validate this and the consensus solution is to get a UPS--seriously!?! If my tone sounds a little peeved, it's because I am--this is table-stakes error/recovery handling. My entire house is down right now and I'm 100s of miles away and unable to do a thing... I don't even know if my doors are closed.
I've enabled remote dashboard--usually works like a charm... until a power outage. I've enabled and paid for Remote admin--no idea if that works because I tried paying for that only this evening to correct the repeat occurrence I describe above... no help. Any by repeat, I believe I've had the C8 for 6 weeks or so now and I'm at 12 brownouts (it's Florida, what can I say) and the C8 does it every time. But now I'm not at home... c'mon @Hubitat_Staff, this needs to be resolved.
I read one post that mentioned enabling some devices for 'mesh' compatibility would cause this. I have (I'm guessing because I can't see anything right now) 70+ devices. I might have enabled 'mesh' on some of them before I knew it wasn't relevant to me... doubt it, but can't tell right now.
I'm a 99% advocate of Hubitat but it's fault tolerance is poor to say the least. Please, please make an effort to resolve this one table-stakes behavior that stains an otherwise near-perfect home automation hub.
There are a number of possibilities here, including that your network comes up in an odd state after a power outage. This situation, power being unexpectedly removed, is impossible for the hub to deal with, because once it happens, the hub is gone. As for when power is restored, the hub looks for its DHCP address. If it doesn't get one, then it's incommunicado. It could be the router is slower to come up upon power restoration than the hub. This is just one of many possibilities, and it's unrealistic to think this problem can be completely solved by software. Software needs power, the hub needs a network to talk to, the network needs power. Take the power away, that's a bad situation.
Yes, this is the best solution for this.
If you have frequenct power outages then the sensible thing to do is use a UPS for the hub. This can be done with a relatively cheap battery, available on Amazon for around $30. These types of batteries are usually used as phone chargers, or deeper phone batteries. The key is to get one that maintains its output even when its power supply turns off. I have one of these that will power a hub for about 20 hours, certainly longer than any typical power outage. I don't use it for my hub as we don't have power outages where I live. But if I did, this would absolutely be the go-to solution.
Yes, seriously. Your hub is a little computer that does a ton of read/write operations to an H2 database. Power interruptions in the middle of writes will call cause database corruption. So a UPS is a good start. A better solution would be to gracefully (and automatically) shut your hub down when the UPS has less than 5% battery power left.
In 4+ years using Hubitat, I've never encountered the issues you've encountered. But then, both my hubs are on a UPS and I do monitor the % battery-life remaining on the UPS while power is out.
The #1 way to prevent problems with your Hubitat hub, your ISP's modem, your router, and your desktop computers: Buy a relatively inexpensive UPS and put those devices on it. If you need more than one UPS to accomplish this, buy more than one. If you have other devices that behave oddly when the power goes out (or browns out), put those on a UPS as well. People think their utility power is constant. It's far from that, and we can probably expect our aging electrical grid to decrease in reliability (a different topic). But if you want to protect your own equipment, the UPS is your friend. This is not likely a Hubitat problem - it's far more likely a problem with your electrical supply.
I hear ya on the startup sequence. However, the rest of the house (87 Wi-Fi devices in total per the Orbi) are all up and running. The hub is technically attached to a UPS since it's powered by the always-on USB port of a Windows Server laptop that runs the house's network infra (DHCP, DNS, VPN endpoint, etc.). The C8's DHCP reservation is also marked as 'in use' and the lease time is a few mins after when power was restored so the hub's DHCP client clearly has sufficient retry logic and obtained an address. While I was NOT unable to connect to the hub yesterday remotely or using the custom dashboard, it was nonetheless working since I was able to control all of my Z-Wave devices using Alexa as a proxy. However, this morning, they're all marked as 'unreachable/offline'.
Do you guys have a way to gracefully reboot it if I supply what seems to be referred to as its MAC address... do I?
H2 databases adhere to ACID and all writes are atomic. The engine uses logging/transactional replay to preserve integrity just like any modern, credible database engine: Advanced. While the C8's implementation may optimize for performance for the in-memory database and, hence, any given write transaction in-flight during a power outage may not prove durable, the database remains nonetheless integral.
In my opinion, the C8 should be able to be treated by its customers as an 'appliance' in this regard.
USB 3.1 always-on (type C) in a Lenovo ThinkPad. Not sure why that matters, though--I know for a fact the hub went down when the laptop did. The power outage lasted 9-hours while a Generac propane generator was being installed. My family and I decided to visit relatives a few hundred miles away rather than melt without any air conditioning... hence my absence.
Well, I don’t want to argue with you, but what you’re asking is a bit much if you want the hub to be tolerant of a wide range of power fluctuations and remain anywhere near its current price point. I’ve had my hub (four of them, actually, in four different buildings) reboot just fine, many times, after just cleanly cutting the power and leaving it off for several seconds. But they perform much better and more reliably in real world situations (rapid on-off cycling as in a storm, or brownouts) on a UPS. If you were having a generator installed, I suspect your house experienced several on-off cycles, and a fair amount of dirty power. And you were powering it through a laptop as well, during all that? A bit much to ask of any device, especially one designed for the home automation market.
We don't have a lot of electrical issues here, but I have a whole-home generator just in case. Nevertheless, I still have two separate UPSs ensuring that all of my infrastructure (PC's, Hubs, Router, modem, switches, access points - and of course the Hubtat hub), stays up during the power gap between an outage and generator start-up . Not only do the UPSs keep the infrastructure running, they also protect all of it during brown-outs, spikes and other electrical issues.
Lastly, I also have a separate monitored security system l that's connected via both wifi and cellular. If my LAN and wifi don't come back up - making the Hubitat hub unavailable, I can reboot all of the infrastructure and the Hubitat Hub through z-wave switches that are attached to the security hub via the cellular connection.
Yup. It's a FrontPoint system (monitored by Alarm.com). Its z-wave capability is so it can act as a smart home hub controlling lights, door locks, thermostats etc. But since I use Hubitat for those things, I only deployed three z-wave switches to force-reboot Hubitat, two internal and four external security cameras and all LAN and wiFi infrastructure devices when needed.
@Madcodger: You and I share reasoning and conclude the opposite wrt/ the C8’s price:resiliency… to each their own.
@lflorack1: We have a ton of power outages hence the Generac but it’s a standby system, not inline so the UPSs remain.
My Flic hub reboots perfectly; same for the Blink storage module. Like these two, the C8 is solid state as are my smart thermostats—all survive and reboot perfectly and all are similarly priced.
On the commercial power front, nod—agreed. Florida’s stormy/electrical climate wreaks havoc with my smart switches, too, so I’m looking for some kind of solution for that as well… in-line power backup seems unlikely without a power wall or the like so I’m hoping to find a way to stretch brownouts from ms to multiple seconds since this has been the primary cause of failures.
Lastly, very cool what you’ve done with the cellular backup. Did you document your setup anywhere that I could take a read through?
As was mentioned earlier, after a power outage and the subsequent power-up, the hub looks for its IP address. If the router or other infrastructure hardware takes longer than the hub to reboot, the hub will not be able to communicate. There are other possibilities, but this seems to be the most likely candidate.
Honestly, the real solution is to add a UPS to protect at least the hub from this happening. Adding UPS protection to the rest of the infrastructure hardware will also be of great help for not only this scenario but to also protect that equipment as well.
