This has been a frustrating week for me. My hubitat was running beautifully for six months. I would occasionally check it to see if updates needed installing , ectc. and no problems--until two weeks ago. Then everything went haywire.
I did a soft reset, and I was able to rebuild everything from scratch--making some minor improvements along the way. And this morning, it suddenly stopped working. I was exploring using the home assistance app to see if it would connect to my hoobs device because there was one little device on the hoobs i wanted on my hubitat--a two button lutron switchc. I could not figure out how to make the token, so I gave up on it. Then I used the package manager to remove the package. I like to keep a clean house without a lot of unused apps cluttering up things.
Then it died.
I did a soft reset using the 8081 port. The dialog box said the soft reset was successful. Then I tried to open the web interface to the hubitat. "Corrupt Database." I did a second soft reset. "Corrupt Database." I used the diagnostic tool to install a previous version of the OS. Did a third soft reset. "Corrupt Database."
I have a backup I created this morning, so all I need to do it to get this running, and then recover the database--created before I started tinkering around with the home assistant app.
This device was installed only last Spring. It is a C7 hub.
I have a similar situation and it appears that every historic recovery I do relegates the oldest backup from the list so now I have 5 backups that all yield corrupt database.
So Iโm going to have to rebuild manually I think.
Would it not be possible to only erase old backups if a new restore actually worked so that this situation canโt occur? I.e. detect that the database isnโt still corrupt?
Rebuilding manually is a pretty radical step. I'd suggest one or two things first. In a different thread the support team recommended a full power down after a soft reset in case the file system was locked. You can also do a "start from scratch" and then do a restore from backkup after the UI is up and running.
If you do a soft reset and a power cycle and then it STILL comes up with a corrupt database I think it's time to reach out to support and have them look at the engineering logs to see what's up.
To Brad, go to settings>backups and you will a table of five local backups. In the last column on each row is a link to delete the backup. Just click on that.
To all.
I tried another soft reset and I would have tried the power doen option after the reset, but the soft reset failed giving me a corrupt database error before even beginning the reset.
How can I reach out to support to have them to look at this. Can they do this remotely?
I tried to force a power cycle by unplugging the USB power supply, and then plugging it back in again.
The situation has gotten even worse. The WEB UI fails to connect, and when I try the 8081 port, I get "Error:failed to fetch." This happens after I paste in the MAX Address in the login dialog for the diagnostic tool. The page comes up, showing the version, model revisiion number and all. However, none of the buttons--including the "soft reset" button are enabled.
@user1067 Yeah this is pretty weird. Unplug power for 5 mins not just in and out. Then boot up. Regardless of the corrupt database message. Now do a soft reset from 8081. See how far you get now. I'm sure someone from @support_team will get to you quickly.
Ok. I unplugged the device, let the dogs out, and plugged it back in. Still nothing. A few hours later, I let the dogs out, and the intruder alert on the HSM went off. I looked at the hubitat UI on the web, and it was running. Unfortunately, it loaded the newest backup from the saved local backups which was made overnight--when the HSM was armed for night. So when I opened the door, naturally the alert triggered and my two siren devices went off.
Then I tried to do a soft reset, and the soft reset failed to wipe out the database. And it had loaded a backup different from the one I loaded the last time I had a successful soft reset. And I cannot reboot the device without it going offline for several hours, and then magically coming back--using the local backup. The autommations seem to be running fine right now, so I do not want to mess with this device anymore.
I cannot control some of the lights in my home without this hub, so I am going to leave it there for now. In the meantime, I ordered another c7 device off Amazon, and purchased the four year warranty plan. Even though I have Prime, it will still take two weeks to get here. I can still cancel the order if I solve this problem.
Now you mentioned "engineering logs." I am a software engineer. Is there a way I can look at these logs myself?
It looks like the hub loses power and reboots very frequently.
Can you try replacing a power adapter with a different one? The power draw is low, so pretty much any USB adapter will work. A different USB cable would be the next thing to try.
You can see how often the hub reboots by checking Logs page, Hub events tab.
The update, reboot, and initialization went smoothly. Previously, I observed that sometimes the initialization I updated to revision 138 successfully. The reboot went smoothly, and the initialization went smoothly. I noticed before that the initialization would sometimes restart and take a long time to complete. This time, no steps were repeated and it completed in an orderly manner.
I was having trouble maintaining a connection to a z-wave lock in the garage, so I was moving the hub around to try to make that happen. And that using a new adaptor since the one that I was using before was built into the powerstrip. It just never occured to me that this would cause so much trouble in other ways. Ah well...
I arose to this lovely event today. Hub logging ceased around 0800.
Pulled power and installed a higher output (1.55A) wart. It did appear the logs showed frequent restarts. Pretty sure I was using the brick that came with it @1A.
Started it to the same error and did a soft reset then restore 2nd latest DB from the hub.
Everything seems OK but I'm going to do a few reboots, shutdowns, hard power pull/cycles.