Corrupt database found - any advise?

[sburke781]

What do you see if you access the hub on port 8081? http : // :8081 Can you access this? Can you reboot from this page? What do you see after a reboot?

Admittedly buying time for @bobbyD or HE staff to assist...

Sounds like you are in a similar situation to @k2.mathew

[alexcapone]

@sburke781

http : // :8081

I can get into the Diagnostic Tool via port 8081:

[sburke781]

What if you reboot?

[thebearmay]

What seemed to be working last night for a couple people that had this issue was re-attempting the soft reset. (Reboot was getting stuck at around 15% or so...)

[sburke781]

Yeah, now that you mention it, I did see @aaiyar suggesting this earlier... to @azz710

[alexcapone]

Yes reboot is successful! And it looks like all my devices, apps, and automations are back to where they were before.

Just a simple question: If my database is curropted and I take a backup afterward and then restore that backup after a soft reset aren't I loading back a curropted backup?

[sburke781]

I think it is ok... but HE staff can answer that with more authority than me...

Glad to hear you are back up and running

[alexcapone]

oh okay. Hopefully the support team can chime in on that.

Thanks to you and @thebearmay for the fast responses! I was starting to have a heart attack this morning.

[sburke781]

Have seen a thread from last month where HE staff were suggesting to keep an eye on the logs following the soft reset to make sure nothing re-occurs that may have caused your issue. Probably a good idea to keep an eye out for anything strange.

[NOT SOLVED] Corrupt database

There are so many moving parts that are literally making changes on your behalf every second. A user may not need to make physical changes, but an brief power outage at the wrong time, an app, a device, a lan or cloud connection can spin out of control for a week and that would be enough to put stress on the live database. With that being said, a Soft Reset will restore a functional database. But it is critical for the user to screen the logs after the reset to make sure that what caused the …

[oldcomputerwiz]

@alexcapone

Just a simple question: If my database is curropted and I take a backup afterward and then restore that backup after a soft reset aren't I loading back a curropted backup?

Backing up and restoring the data base removes the corruption. I have had bad backups though and had to load a prior backup. A good reason to keep multiple backups on another device.

[sburke781]

Good suggestion, I download 14 days worth to a raspberry pi. From memory there is a reasonable number kept in cloud for you as well...

[gopher.ny]

@alexcapone

If my database is curropted and I take a backup afterward and then restore that backup after a soft reset aren't I loading back a corrupted backup?

It depends on what exactly got corrupted. Generally, I'd go with a previous backup.

[alexcapone]

How do I know what's been corrupted?

[Sebastien]

@alexcapone

Another question, does a backup contain all apps and automations? In other words, will it revert exactly as it was before the soft reset?

It contains all apps (including automations and app code), devices and device drivers. It doesn’t include files that are added through the file manager. Only the cloud backup include the z-wave database, but restoring the file backup leaves what’s there alone, so this doesn’t cause any issues.

I have done many soft resets and many restores on my main hub and have not had any issues (yet…). It’s scary the first time, not so much the 5th….

[csteele]

@Sebastien

It’s scary the first time, not so much the 5th….

Seems like so long ago now but I was doing soft resets 2-4 times a day when I needed to test different HubConnect configurations. Soft reset to a backup that has no HubConnect and install it, through soft reset to a config that had the most recent version and do the upgrade.

I haven't needed to do one very often but the practice of doing it 4 times in one day sure improves confidence.

[alexcapone]

After doing a soft reset I'm starting to see this alert again:

I assumed it was related to the ghost node that I can't find a way to delete. How do I confirm the root cause of this alert?

[thebearmay]

The only time I had a DB size issue I was also fighting a ghost node. May not have been related, but it sure felt like it. That being said, what are your events and states set to, and how do the device run time stats compare to that setting?

[alexcapone]

So events and states were set to 100 which I think is the default. I changed to 11 after reading this:

Alert - Your database is growing

Just set the value to 11, no need to reboot right away. It should, at a minimum, stabilize things.

There are certainly a large db log:

Hopefully this stabalizes after going from 100->11 for event/state history

[neonturbo]

Things like weather, TVs, Roku, and things like that have lots of events, and they seem to create (falsely or not?) that database warning. Not picking on any particular app or device or driver here, but those are often very intensive things with lots of events in general and it can't be helped that they need lots of storage.

But storing needless events seems to make things sluggish and the hub run worse. Setting to 11, or even less for some apps, seems to really calm things down. In particular, I set weather to 1. Why do I need to keep days upon days of past weather events? Or TV stuff, isn't one or two of each event enough?

Before they gave us the ability to manage events in one of the last couple updates, I had hundreds upon hundreds of motion sensors events stored from a year previous! That was ridiculous to say the least. Wasted space at best. Thank goodness we can limit things now.

[alexcapone]

About 24 hours ago I changed both event and states to 11:

Yet my database keeps growing from yesterday:

I can only think it's related to that ghost node but I have no way to prove it.