Updated Vero 4K+ freezing/crashing

A little background…

I have two Veros, one is a 4K and one is a 4K+.
The 4K is connected to my network via Wi-Fi and only has the official remote dongle connected, the 4K+ is connected to my network via ethernet and only has the official remote dongle connected. Both use HDMI to connect directly to a TV. Nothing else is connected to either Vero. The 4K+ is connected to a LG 4K TV.
Both are using a shared database on a Windows PC using MySQL and both were running smoothly with this setup on the November 2020 release.

I recently updated both Veros to the latest August update and have run into a problem on the 4K+.
The system as a whole works without issue while in use, but after some time it becomes completely unresponsive. When looking at the Vero display on the TV the screensaver is on by this time (dim). There is no response to the remote or trying to login via FTP or SSH, also the official iOS Kodi remote connection times out.
The only way to “fix” it is to remove the power and power on again. It’ll then run fine again until it freezes again. There doesn’t seem to be any specific time until this happens, I just check it occasionally and sometimes find it frozen.
I had a look at the CPU usage on the HMI on both Veros and the 4K+ seems to be running all four cores at >80% even when idle, i.e., just looking at the CPU usage. The 4K seems to idle around 15% across cores.
The odd thing is the 4K does not freeze/crash since the update.
Any help would be appreciated.
Here’s a link to the logs from the 4K+. https://paste.osmc.tv/tunovaxumi
I enabled debug logging, rebooted twice, and left it idle until it froze, I then had to cycle power to grab the logs.

What does the clock in top right corner shows at that time? The current time or a older time?

Seems to get network issues at one point in time

2021-08-14 18:51:00.702 T:7759    ERROR <general>: Unable to open database: MyVideos119 [2026](SSL connection error: A packet with illegal or unsupported version was received.)
2021-08-14 18:51:06.671 T:7724    ERROR <general>: Unable to open database: MyVideos119 [2013](Lost connection to MySQL server at 'sending authentication information', system error: 32)

It shows what ever time it froze at and the debug info on screen is also frozen.
The 4K doesn’t seem to have the issue connecting to the same database, but to be honest I have no idea what those MySQL errors even mean.
Since that log was collected I did try switching the kodi user to mysql_native_password from caching_sha2_password but that made no difference to the freezing status of either Vero.

That seems to be a similar issue as reported here (while that is on a Pi)

The MySQL issue may not be directly related could be just that network fails when the Vero freezes.

If the network fails, the device might not be freezing but rather Kodi is hanging and not responding gracefully.

The output of top would be interesting

TIme is stopped so interface froozen

How should he run top when network failed :wink:

I’m talking about finding out why the CPU use is so high when idling. He must have had console access at the time.

Ok, must have misread that. Thought that was remotely before the freeze. but yes than top make sense

I don’t think the network is falling over. Both Veros have been connected to the same router for some time and it has a few other devices connected to it via ethernet and Wi-Fi (the MySQL server included).

Even if it was network related why would there be no response from the Vero remote?

When I referred to the CPU usage I was looking at the page/screen on the Kodi interface on the TV not through any console access, and I’m afraid I don’t know what “top” is.

You could connect to the Vero via ssh and run top in the console then you would see CPU usage and might also see the crash.

Details regarding how to access the command line interface can be found here on our Wiki: Accessing the command line - General - OSMC

Just FYI RE: CPU load during the freeze, for my issue: Frontend freezing since 19.1 update didn’t seem to cause a CPU spike, i have data from the last week every 15mins and there was no continuous high usage after the freeze.

My issue was fixed by removing a repo that wasn’t responding, this was most likely a new Kodi 19.1/OSMC 2021.8-1 issue however as i assume the repo was broken before the upgrade.

Thanks, I’ve managed to connect via ssh and used the top command. I get some rusts from both Vero boxes. How would I go about copying that info here?

I’m currently travelling for work so as a temporary measure I’ve swapped the location of the two boxes to see anything changes. The 4K is now on Ethernet and the 4K+ is now on WiFi.
Interestingly, 4K CPU usage remains low on Ethernet and 4K+ is still high on WiFi.
Both boxes “should“ be configured identically because I used the OSMC backup from the 4K to setup the 4K+.
I’m sure I’ll hear from home the instant either one falls over. Will add more info when I can get it.

Thanks. The high CPU usage was not actually pre-freeze it was just while using the 4K+ “normally”. Freezing happens much later when the box isn’t being used. I’m tempted to wipe the 4K+ and start fresh, but I’ll see about that when I’m back home.

Well could add a screenshot. Basic question is what is the CPU usage and is it all used by Kodi/Mediacenter?

Also doesn’t harm to upload new logs after the swap.

Hello everyone, I thought I’d give a little update on this.

So… I did a fresh install with the OSMC Installer on both boxes using the 2021.08-1 release.
I also reverted MySQL back to 5.5.62 and restored the sqldumps.

Both boxes are running fine at the moment, apart from one minor niggle where the GUI occasionally flickers green (the V4K+ connected to a LG 4K TV), this happens on the OSMC skin and Estuary. It happens randomly and it’s just a quick flash of green.

A couple things I noticed during the reinstall:

The GUI defaults to 1080i - easily changed.
The default setting of the CEC allows my TV to put the Vero into standby - changed in the CEC peripheral settings.
The OSMC backup functionality keeps throwing an error about the folder location - fixed after a couple reboots.