Lost Hub configuration, reseted to factory default!

Yup. I ordered my Smarthings from Amazon USA.
Anyway I suspect one of their databases got corrupted and is replicating the corruption over their cloud.
I’ll explain how I came to this conclusion.

As I did not have any additional location (it was originally using Home before it went wonky, and location was named ‘Home’ again after it went wonky). I figured I would add and configure my devices again. However smarthings could not detect anything!

So here’s what I did, I deleted the location ‘Home’ and created a new one called ‘OliHome’.
I then ran a detection again. This time it detected all my devices!
So i figured okay. Just spend a few hours setting everything up again.

But there was a few times, (in IDE) when I clicked the links for my Device Handlers and SmartApps. I would see my old device handlers and My Smartapps there.

Then if I log off and log on again, they would be gone. 9/10 times I would not see them but those other times…
This got me thinking one of the servers lost my data but the others had my old data in the cloud.

Also might explain why smarthings didn’t detect any of my devices when location called ‘Home’ was created (after it went wonky). Because somewhere in the cloud, my devices were already detected so it wasn’t detecting anything new. It was only after I deleted ‘Home’ and created a new location name did it start working.

Anyway, I’ve reconfigured my smartapps again (with some new changes) and everything looks good. (spend nearly a day fixing it all up). I now hope my old data doesn’t return and my new one doesnt get wiped off again.

Ok. The reason I said about the server address is people with uk hubs have been caught out using the wrong url when looking at their ide.
I still think @tslagle13 that the server our hubs are connected to should be shown somewhere in the hub.

We have confirmed this is a trending issue and are working to address the root cause. It appears that the cloud platform is informing the mobile client that the account has no locations and is forcing you to reset the Hub. This appears to only be affecting a subset of Android users on our NA01 shard. When it occurs, force quit and re-open the mobile app. Please DO NOT run the initial getting started walk-though or you will lose your connected devices.

We are working on a fix now and will provide an update with more information ASAP.

Worth noting I saw the same thing for some hours earlier this week on IOS 10.1

Thanks for the heads up! We are seeing similar symptoms, but different failures on iOS. We identified the cause and are working on a fix. Status update coming shortly

1 Like

Hi @Aaron - I’m on the NA02 shard and experience the same issue (Android 7.1). Let me know if you need more info.

Is it confirmed that the hub config is wiped after proceeding with the welcome/activation key input?

Any other workaround to recover the zigbee pairing?

OR i should move on and start the salvage process?

Pls consider creating a backup feature to backup critical hub config. This is a terrible failure mode to recover from.

Imagine if this failure occurred to the masses and more customers lost their config and have to restart from scratch.

After this episode, i’m hesitant to recommed smarthings to my non geeky friends. It’s still too bleeding edge stuff.

1 Like

This is probably the most important impactful issue I have seen. I am sure there are folks on these boards that would talk about their setup and how simple and perfected it is. I would like to know how that avoids this?

Anyway. It doesn’t. There is a lot of Victim blaming in that regard.

Holy crap st.

1 Like

I agree.

I freaked in empathy when I read the Incident Report.

I wonder how many Customers affected??? I can certainly image that a non-trivial proportion of Customers who were asked for their Welcome Codes just went ahead and input it.

1 Like

Support tried their best. My hub config is toast.

Removing all smartapps and devices and starting over.

This time round, I’m pairing my china Kudled led fixtures to hue hub instead.

Other sensors and switches will still goto ST hub.

Still waiting for ST to reply on how to prevent a repeat of this epiaode…

Suggestions as follows:

  1. Provide hub backup and restore feature
  2. Warn and double warn before proceeding to claim a used hub and wipe existng config.
1 Like

I’ve rebuilt from the ground up once already. I would not do it again. It took a month’s worth of my time.

At this point, HA is not providing a value - it’s more of a burden.

If SmartThings were able to develop this ability, then they would have already delivered the “Hub Migration Tool” promised for the end of 2015.

Someday it will happen… Or not.

Calling routines/hello home actions from button controller seems to be broken.

Sigh.

Commanding on/off of bulbs from button controller seems ok.

For a bit more detail on this… There was a bug (it’s probably existed for a while and just surfaced) where the platform would return no locations (generally when under stress). We put in limits server-side to mitigate the vast majority of possibilities where this could occur (also - as a separate workflow - the recent database work helped alleviate the stress). We will also be adding updates to the next mobile builds as a secondary prevention measure.

IF you come across an unexpected getting started screen again, force quit and reopen the app. Also shoot a note to support@ with the approximate time it occurred so we can investigate.

Aaron -

Wouldn’t a more solid mitigation be called for until you can be sure it doesn’t reoccur?

Like preventing registering using the new device reg code, or whatever it is, without manually emailing support first so they can verify some things?

It just seems like such an extreme penalty. Now that I know about it, I expect I won’t fall for this issue - but despite the emails, the status.smarthings.com, etc I can see joe consumer still getting caught up in this unless it is completely resolved.

1 Like

Good god, yes. I understand platform issues are harder to fix so why not immediately add checks to prevent this from happening.

This is EXACTLY what led me to dump Lowe’s Iris. A similar bug existed in their app when removing devices where the UI would lose reference to the device being deleted leading the user to remove the hub and all devices. Instead of helping or acknowledging the bug, support literally told me to be more careful the next time.

They have since fixed the issue, but it cost them their largest user and a lot of bad PR on the blog.

1 Like

A few more data points to contribute.

  1. Of course I killed the app/process on my android phone in an attempt to recover during the initial hiccup on Monday. It didn’t WORK despite multiple attempts.

  2. Then I cleared data and restarted the app. Still no go.

  3. I left it alone as I thought it was an app/platform problem.

When I reached home, NONE of my sensor triggered smartapps fired?!!?

So I thought it must be deeper than just an Android app problem.

Wife was demused to see the wonderful HA contraption all falling apart.

That was when I went to dig out the box to get the welcome code.

So yeah, it went downhill from there.

Question: Given than my smartapps didn’t fire when I reached home, was there another secondary issue at play?

1 Like

Looks like this happened to me. Wife’s iPhone wasn’t able to login to smartthings while I was traveling and asked for the welcome code. Today I entered it and a second location was made which kicked the hub off my android (became inactive). Found this thread and removed the location from her account. I am able to see the hub is now online, however, none of the zwave or ZigBee devices work. I can still see them via the online portal, and they appear in the app, but they don’t function at all. Tried removing a light switch and I can’t pair it. But the ecobee, hue and alexa appear to be functioning in the app. Tried running the zwave network clean via the app and it immediately says complete.

Really disappointed by this, I’ve invested a ton of money and to have it all fall apart this easily is pretty jaw dropping. Really hope something can be done. Otherwise, I’m not sure this is a viable platform to base HA on going forward, especially as the integration becomes more complex with all the various devices and software. Just really at a loss for words.

2 Likes

Eh? I tot the bug had been patched over at the server side?

The only way is to bite the bullet.

Delete all old ZigBee and zwave devices and related automation n start over.

Unfortunately, st is the only game in town that is 3rd party friendly. So… That’s the price for being bleeding edge.

Its definately not consumer friendly.

Over the Thanksgiving holiday, I setup a V2 hub for my sister. I just received a call tonight from her indicating she was adding her husband (via an invite) to ST and suddenly lost all the configs. I wasn’t home when she tried to contact me so she contacted support and the long and short is we now have two accounts and the z-wave device is no longer active (inactive on the old account and not seen on the new account).

So here are my immediate questions:

  1. Does the hub still have the pairing for the z-wave devices ?
  2. Rather than start over, can I simply (via the developer portal) force an interview of the devices ?
  3. Anything other than starting clean that I’m not thinking of ?

From the many threads I’ve scanned, it looks grim at this point so I’m trying to be optimistic in helping her from such a great distance.

Much appreciated,
Cuda