Went into reboot loop unexpectedly

My RPI 3/OSMC system has been running great for quite some time. In fact, I watched some recorded videos last night. Then, this morning I came downstairs to catch a show while I was eating breakfast and I found my PVR in a constant reboot. If do a restart it goes thru, what now appears to be, a partial boot up sequence on screen (gets to the blue screen with the OSMC logo, goes a bit further) and then it shows the “sad face”, goes dark for a few moments and then back to sad face, and continues to recycle…my sad face → :frowning:

Idk for sure, but there may have been a brown out last night (clocks are OK, tho).

I can get in thru ssh just fine (well, mostly…once it did a reboot and kicked me out. As I write this it is in a constant reboot mode).

I did a " tail -F /home/osmc/.kodi/temp/kodi.log" to see the boot process and since I couldn’t remember if I should post it here, I put it on paste.bin:

One thing I noticed right off is is there is a CProfileManager that can’ seem to open a //masterprofile/profiles.xml and that’s the point where the system resets.

Definitely open to any insight as to why and how I can correct the problem.

Thx iin advance…cheers…

My first guess would be that your SD card may be failing. You could try saving the kodi configuration and try with a new clean setup.

sudo systemctl stop mediacenter
mv .kodi .kodi.save
sudo systemctl start mediacenter

And see if Kodi works again. You could then start restoring your setup.

Interesting, indeed.

Does that mean I need a new SD card? Also, does it make any difference that I have the system set up with a hard drive (went thru the process of moving the system to it way back when)? What does that mean for what’s stored on the hard drive?

Thx…

It would have helped if you had mentioned this! Without proper logs it’s only a guess what’s happening.

So you actually have OSMC installed on a hard drive?

If you can SSH in, try posting logs manually

grab-logs -A

and share the link

My bad. I realized it after I posted my original :frowning:

I did the mv .kodi .kodi.save command and I am now out of the reboot cycle…however, now the user interface is nothing like what I had set up. is that the restoring you were referring to?

Can I do a “mv .kodi.save .kodi” and get back my configuration, is that going to put me back where I was?

And do you still want to see the logs?

Cheers…

It will get you right back into the Kodi restart loop (it’s not rebooting!)

But it would help to restore the original again, and do the grab-logs so we can see what’s causing Kodi to crash.

You didn’t indiate which PVR client you are using but it sounds very much like the issue I had last night / today with the unofficial HDHomeRun PVR client. See this thread. . If this is the PVR client you are running then the fix is:

  1. SSH into OSMC and go to .kodi/addons/pvr.hdhomerundvr/
  2. Sudo rm - r *
  3. Reboot
  4. Reinstall from Zip file version 2.0.3

This will retain your PVR addon settings.

I restored .kodi but reboot loop is gone. However, still have default UI.

Here’s the results of “grab-logs -A” after a reboot.

https://paste.osmc.tv/gasuquyata

Let me know if anything stands out.

Cheers and thx…

Since the problem is gone, it will be hard to guess what happened.

Just install the skin you were using before and see if everything is back to normal.

Scratch that. Turns out the .kodi.save ended up in the wrong place. I redid it and now I’m back to the problem.

Here’s the new results for the logs:

https://paste.osmc.tv/fusobukuta

It’s failing to open the profiles.xml. Did you save the .kodi directory that was working? If you did you could try copying that back into the original saved directory.

The real question now is why did the profiles.xml get screwed up? Do you have a linux system that you could connect the drive to and run an fsck on it?

Feels a bit like we’e in a time lag with each other…

I did the mv .kodi .kodi.save and .kodi.save was moved under .kodi so the system must have gone back to default (noticed all my addons were then gone from .kodi). I then moved .kodi.save back up 1 level and renamed it back to .kodi (1st renaming .kodi.save to .kodi.orig). Then it was back to the problem.

In my 1st posting where I did the “tail…” I saw the same problem about failure to open profiles.xml. Is there anyway to try and restore most, if not, all of it somehow?

Yes, I could move it to my Fedora 30 box, but that would be a pain. No way I can do an fsck on it while in the RPI 2/OSMC system?

Just noticed something else. Not sure if it’s the same one referenced in the logs…there’s a profiles.xml under.kodi/userdata and it shows 0 bytes and when I open it it’s empty. May explain why it can’t be “opened”. The one located under .kodi.orig (which is what I named the .kodi that showed up after I did the mv… and “start”) has about 1KB and does have xml text in it.

Seems like my orig profiles.xml got mucked up somehow. Wonder if a copy exists somewhere that I can use to restore it?

Not unless you’ve done backups. But you could use that one as a starting point to get going again.

Not that I actually remember :frowning: I will going forward.

It means a lot of work to get the system back to where it was. Or, can I move everything (like addons, userdata,etc.) except the bad profiles.xml over to the new .kodi and avoid a lot of work? Or is that a disaster waiting to happen?

Or, can I simply keep the .kodi with the bad profiles.xml and replace the bad one with the new, good one?

You could try that. If that doesn’t work, then you could try restoring things by copying the directories.

I copied over the new, good profiles.xml to the orig .kodi directory. Everything is pretty much back to normal. At first glance, the only things that are not back the way I had them are changes to the main and submenus I had made. I expect I may find other settings that have to be redone, but it feels like it’s 90% or more there. Lucked out.

As soon as I have things back where I want them, I’ll make a copy of the .kodi directory, just in case. I also plan to get more committed to backups.

Thanks much for your time and help. I truly appreciate it!!

Cheers…

I would suggest that you shut down the Pi, move the drive to your Fedora system and give it a good test to be sure that it’s not failing so you don’t have this problem again.

Is the drive self powered?

@jeff2

Didn’t mean to ignore you.

I am using the HDHomeRunPVR client, but if you’ve followed this thread it turned out something glitched my profiles.xml file (basically zeroed it :frowning: so now I have things back to a more normal state. Still some work to do tho.

Thx for the suggestion. I’ll keep it in mind next time something similar happens and it’s not my profiles.xml file :slight_smile:

Cheers…

That would prob’ly be the smart thing to do. When I get the chance I will do exactly that.

It’s not self-powered and since it’s a 3-1/2" HDD it requires external power to run it from a USB port (I use an active USB-SATA interface cable to connect the drive since that’s how the RPI 3 rolls with HDDs). So, I can do it, but it’s a pain. It might be just as easy to actually do the testing “in system”.