April update causes System freezes

Help and Support Raspberry Pi
61

Same Issue.

RPi1 with latest OSMC Update. MySQL Database (advancedconfig.xml)
Exactly same config worked for month ago. Problem occured after April update.

62

@Sir_Mc_Tod (and others with OMX errors in log) please update to May release and report back.
Post a new log if problem persists.

If you didnā€™t have OMX errors in log, then update and start a new post (with logs) if there are still issues.

63

Can you create a new thread? Your issue doesnā€™t seem related to original posterā€™s issue (OMX errors on log).

64

Hi,
the issue is exactly as described in the first post of this thread. I can create a new one, but I could copy-paste the same text and the same title

65

Yes, please do. The underlying issues (and so the fixes needed) are not related. We are more likely to get solutions if different issues are not mixed up.

Iā€™d like this thread to be for OMX errors in log, like the original poster had.

1 Like
66

OMX errors as before post May update. So issues not resolved. Affecting Pi1 B+ & Pi2.

My Setup is 6 x piā€™s MySQL DB, media shared over SMB (No SMB issues).

All other clients unaffected, Android tablets and phones, ubuntu etc.

Cheers
Spart

67

@sparticle
please post a debug log

68

So what would be the command grab-logs -A ? passwords are clearly visible in these logs and are publicly visible and are not obfuscated or redacted!

Cheers
Spart

69

You can edit the logs and run paste-log on redacted files manually

You can always just PM @popcornmix the link as well. I am the only one with access to Pastes without a URL so the paste would be safe (the URLs are hard to guess)

70

I opend a new thread beacuse I think my error is also not the ā€œOMXā€ error.

You can find it here with all Logs:

@popcornmix

71

Here you go. http://paste.osmc.io/ufozegagov

Cheers
Spart

72

I just updated.
Problem isnā€™t fixed :frowning2:
http://paste.osmc.io/golakekufe

73

@popcornmix

Do you know what version is safe to regress to. Rather use an older version that works as the grief level is high. This issue is making the piā€™s unusable as they have to be either rebooted or sshā€™d in to reset them. They only last for a short period then the issue appears again!

Cheers
Spart

74

Your issue is not related to original posters.

18:35:47 15776.427734 T:2952786928 ERROR: COMXCoreComponent::DecoderEventHandler OMX.broadcom.image_encode - OMX_ErrorInsufficientResources, insufficient resources
shows we are out of gpu memory. Can you try increasing it to gpu_mem=256? (It is currently 144) to see if that is enough?
Also make sure ā€œenable higher colour depth artworkā€ is disabled (in system/video settings).
Are you using the default osmc skin?

75

Debug isnā€™t enabled from this log. Iā€™m not seeing any errors either, so you donā€™t seem to be hitting the problem from original post. Can you explain exactly what occurred?
Also can you remove the overclock?

76

@popcornmix Wilco. ā€œenable higher colour depth artworkā€ is already off. The 144 setting was the default setting for the B+ Piā€™s that have 512mb.

The log is full of these these:
17:02:17 199.963348 T:2721141744 WARNING: Create - unsupported protocol(image) in image://smb%3a%2f%2fxxxx%3axxxx%40xxxx%2fTV%2fFrom%20Dusk%20Till%20Dawn%20The%20Series%2ffanart.jpg/

And these:
17:14:06 909.117798 T:2687587312 WARNING: COMXImageFile::GetCodingType progressive images not supported by decoder

For different TV shows and movies. If I check the physical files on the NAS share all is well. The same for all other reported .jpg problems. Physical files are all ok.

Default skin for all our machines is Aeon Nox current version 5.7.3 this skin has been our standard skin for a long time.

Where I used to be able to scroll through the Movie list on InfoWall view it is now very slow and a lot of the thumbnails do not render or take a very long time. The text part is displayed instantly but the matching fanart above it take a very long time (lost the will to live amount) to catch up. It was not like this before April. Piā€™s were very responsive in all menus and media views and played media pretty much instantly from the NAS.

Thank you for taking the time to have a look at this issue. If you need me to do any further experimentation or testing please let me know. I will report back on the memory change.

Cheers
Spart

77

@sparticle
Progressive images have never been supported by harwdare decoder. They always need to be decoded by the arm, so that is not an issue.
Running out of GPU memory means that all GPU jpeg decodes fail and fall back to arm decode, so running slowly would be expected. It also makes it more likely a critical allocation will fail which may cause hangs or other unexpected behaviour.
Start with 256M for gpu_mem and see if that avoids the issue. I suspect a smaller value may be possible, but start with that.
Aeon Nox is rather a heavy skin for Pi 1 - you would probably have fewer issues with confluence or OSMCā€™s default skin.

78

@popcornmix issue not solved with 256 memory. Hang on return to main menu no further output to logs. I have been running with tail -f -n50 ~/.kodi/temp/kodi.log from a ssh session whilst navigating the pi.

It can only be revived by restarting mediacentre with sudo systemctl restart mediacenter or rebooting.

I take your point on the skin but we have been using it with no issues for a long time. Issue still occurs on Confluence skin!

Cheers
Spart

79

@sparticle do you still have OMX_ErrorInsufficientResources in log?

80

@popcornmix will need to enable debug logging again and check. Please bear with me. Currently testing 2 of the piā€™s

Cheers
Spart