I still have this occasionally as well, running 2.3.9 betas. Last time, less than 2 weeks ago. Once more I disabled the automated backups.
I awoke this morning to no zwave on my C8 - it failed after a cloud backup. I install new versions within a day of release - so current is 2.3.9.166. A reboot did not rectify the situation, but a shut-down for 10 minutes appears to have. The last occurrence was a couple of months ago. Is anyone tracking how (in)frequently these occur?
Is there a way to trigger a notification if a backup fails? Thinking Pushover, but could be anything.
I’m using Node-RED to trigger a notification based on the success/failure value in the cloudBackup hub event. I guess you could do the same in RM. Unfortunately, I don’t use RM or HE rules engine, so can’t give an example.
So, just to nudge this, I have a C-8, running 2.4.0.146, and cloud backup failed and z-wave was offline when I woke up (Tuesday?).
For me this happens about every 1-2 months.
Gracefully shutdown and unplug for 15 to 30 seconds seems to restore it, but it's annoying an nor something that can be fixed if I'm not home... 
Are you sure that Zwave was offline, or just assuming because as part of the message that the backup failed it tells you that? The message hasn't been corrected yet, but in most instances (since 2.3.9.xxx IIRC) ZWave has come back online. Still need to do a manual cloud backup to clear the message.
I had this happen on 2.4.0.145 on a C8 recently. Zwave was unresponsive. Saw the alert and shutdown, power cut for 30 seconds, restarted and everything was back to normal. I then initiated a cloud backup. It completed successfully.
My mesh is stable. I saw this last in June of 2024.
My locks and garage door opener are zwave. I am worried this will happen when I am out of town and inconvenience the family. I am considering options such as replacing those devices with thread/wifi/zigbee alternatives, moving them to HA, etc.
Although I take them weekly, I rarely have a cloud backup not suceed. The last 3 that had a Zwave failure (spread over a few months) recovered completely without a power down when I did a manual cloud backup to clear the alert.
Just happened to me last night. Backup failed and zwave offline. Wish that a Hub restart from the UI would reset the radio instead of having to shutdown/unplug/plug
1 Like
That's why until it is 100% fixed I would recommend not doing them automatically at all.
Do the full cloud backups manually after major changes and when you're home and can reset the hub if it messes up.
Either that or come up with an external way of gracefully shutting down your hub, removing power, and reapplying it remotely.
That, or continue to do them automatically, but don't complain if it messes up when you're away from home. 
2 Likes
I agree that the issue needs to be fixed but there really isn't a need to do cloud backups when you aren't at home. Nothing is changing so the backups taken while you are gone are all going to be the same.
Just turn off automatic backups while you are gone and as @JasonJoel suggested, set up a way to remotely power cycle the hub. It's only happened once in my 5 years of HE use but the need to power cycle can arise, and in this case, it wasn't related to cloud backups.
Just know that, while the odds are in your favor the database won't get corrupted, it might happen if the hub isn't shut down gracefully. It's happened to me. So, you'd have to restore from a backup when you got home.
1 Like
Thanks. Been here for 5 years and have had a couple of corrupted databases.
The hub is on a battery backup and I get notified when the power drops and the hub will shut down 2 hours after that. Remote power cycle to bring it back up if the UPS doesn't croak before power is restored.
I only kill it without shutdown if it doesn't respond and it's only happened once. I then immediately did a save and restore.
Now you can just do a database rebuild during a normal reboot and I will do it on occasion just because.
Just try and cover all the bases.
I reluctantly turned off my weekly cloud backups. It’s just one more thing to remember to do manually and will add complexity to a restore if it’s needed.
In my case, I frequently go in and tweak things in rule machine, basic rules, button controller, etc. I also often wait a month or more between updating the hub to make sure I move to a stable version.
It was really nice to know that if I had a big failure, I could restore to a new hub from my cloud backup and be back to a close-to-current state. However, if I haven’t done a platform update in say two months, the cloud backup is old and I’ll have a lot of tweaking to redo. Maybe I could leverage a local backup for that, but as it just adds complexity.
I really liked that feature and it’s a bummer to give it up.
1 Like
Scheduled cloud backups are integral part of Hub Protect service and while it is reliable for most users, there is a small number who are affected by the Z-Wave backup failure. We understand that for these users is not only a major inconvenience, but also makes the service less desirable. Our engineers are aware of this problem and working on a permanent solution to make the Hub Protect service reliable for all.
3 Likes
Because I haven't had any screw-ups for a while, I went back to a weekly (from biweekly) schedule, knock on wood.
I just now changed the time to 9AM, so I'm not wandering around in the dark half asleep, just in case, lol.
Thanks for the update @bobbyD
1 Like
If it was online, it was not working. None of my zwave devices were responding (motion sensors, switches, contact sensors, etc.). That's usually how I know the zwave is offline.
I don't know if zwave causes the backup failure, or vice versa. They are always discovered after the fact.
Any new news on this issue? My Zwave just went offline again, and I had to shut down and unplug the hub again to get it back. Looks like the cloud backup failed at:
cloudBackup Failed steps: Z-Wave FAILED 3/14/2025 2:21:07.343 am
Also, that was the only error in the HUB events for that day. The prior day cloud backup succeeded.
Also, I seem to always see "hub load is severe" after I do this reboot. Eventually, it seems to settle down. Is this normal?
You should consider joining the beta
2 Likes
