Update - it looks like Moode (an audio player) also has this same issue in the more recent builds. I had version 6 and was trying to update to version 8 and am hitting the same roadblock. I'd be sad to think that these fast and (otherwise) high quality µSD cards are not useable with newer Linux builds.
I'm not sure I understand what changed in the underlying OS that now rejects these cards while they worked fine before - these cards that are great performers in all the other uses I have for them...
2 Different RPI3s, Multiple cards (but all SandDisk High Endurance), Both LibreElec and Moode images, 64 bit and 32 bit versions. All fail.
My guess is that this isn't going to get fixed because the "solution" is to use a different card.
Posts by Meestor_X
-
-
Just doing it in the interests of science. Learning is a good thing for no reason other than knowledge.
Meestor_X thanks for the new log, this looks fine now.
The next step will be that you create a bug report on the RPi linux repo here https://github.com/raspberrypi/linux/issues
.
.
.
If you want to proofread or check your bug report before you submit it on the RPi repo then just post a draft of it here in this thread and I'll have a look and comment on it.
so long,
Hias
Copy that. Probably won't have time to do all that until the summer. Maybe it will have miraculously have fixed itself anyway by then.
BTW, I want to use a 256g card because I like to have content without having to have a USB stick or external drive hanging off the device. Another reason I prefer other devices where you can install an internal eMMC or SSD.
I'll just use LE10 for my upcoming trip and check in on the status of LE11 when I get back.
Thanks, everyone!
-
You're saying you installed the latest LE11 on the RPI3 using a 128gb µSD card and it booted without error?
Interesting.
So here's what we know:
- Not a voltage/power supply problem
- Not a card quality problem (The SanDisk "High Endurance" is a high-quality, much-recommended card, and tests perfectly)
- If the bug is in RPiOS, it has an issue with certain µSD cards of a certain size. Cards that work fine with previous versions of RPiOS and other distros.
Where to go from here?
-
It always seemed a bit ridiculous that a micro-USB would be the connector of choice for such a high current draw, but ultimately in my case, the issue with the "low voltage" ended up with the cable, not the power supply. Even a 6v battery powered (almost limitless amperage) power supply didn't fix the problem until I found a very short, thick cable.
But, even after all that, the issue still remains. Sometimes it's not the usual suspects like SD cards and power supplies, but you're not wrong to ask to have those items checked and resolved first!
Hopefully this log will offer some insight into what's going wrong?
-
The PSU is 2A, I don't have any of the "official" RPI power supplies, but I'll see if I can wire up my desktop power supply and give it exactly 5v at any amperage it wants.
How a micro-USB cable or that tiny connector can handle 2A or more is pretty surprising. Those pins are tiny on that connector!
-
Is that the only error you're seeing? No other errors except low voltage?
It'll take some time for me to find another power supply with higher voltage. In my experience, the RPI ALWAYS complains about low voltage no matter what power it's given. Another reason I'm done with RPIs...
-
New power supply and cable. Odd that it would be a problem only with this version of RPiOS?
Would have gotten this up sooner if there wasn't a 40-minute required delay between postings. It's a bit excessive.
-
Whoops, missed that part about the new command needing to be on the same line. Weird that it can't parse whitespace!
Anyway here's the updated output, hopefully this helps.
-
New dmesg.txt
Any way to lower the # of minutes required between posts? 40 minutes seems a little overkill for bot protection...
-
Here's the dmesg.txt file.
-
Thank you, HiassofT , I may give that a try, just in "interests of science"...
5.10 booted up fine.
5.15 did not. Same error I was seeing with LE11.
Where do you suggest I go from here?
BTW, I use the SanDisk Extreme "High Endurance" µSD cards, which are the white ones. They are more expensive, but, for me, they have been 100% reliable and quite quick.
-
Fair enough. Enough time spent on this issue.
-
Ok, TY.
Well, I don't see a single person reporting the same issue I'm having, here or on the RPI forums, so it might be premature for me to report it as a bug. Could be something specific to my RPI3 or the SanDisk cards or LE11 or ??
I would think others will have tried larger µSD cards and if there were a widespread problem, there's be other reports of it. So, I'm a bit confused as to what the actual problem could be...
-
Should I try rpi-update in case my firmware is out of date on this RPI3?
Seems when I ssh into LE, the usual commands don't work (apt update/upgrade, sudo, rpi-update)
-
We have a history of previous LE11 (nightly test ) builds here:
https://test.libreelec.tv/11.0/RPi/RPi2/You could check through them, finding out with which build the issue started might give some clues.
so long,
Hias
Thank you for that link. Even the earliest LE11 doesn't work. The latest that works is LE10.
Before I post a bug report on GitHub, can you tell me what kernel LE11 uses and is it 32 or 64-bit?
-
Sorry, Pete, I know you mean well, and I do understand that µSD cards can be trouble, but this time I'm afraid you're barking up the wrong tree. I do appreciate your efforts in trying to help though.
Please see:
PostRE: LibreElec v11 on RPI3 won't start
TY. I will do that. I assume there's no way to create a LE11 build that uses a slightly older kernel that perhaps doesn't have this bug?
I do like the idea of being able to properly test an SD card, though. Wasn't really up for figuring out how to get f3 installed on my Mac so I took the lazy route: https://github.com/vrunkel/F3XSwift
Tested my cards and they all pass.Meestor_XMarch 14, 2023 at 10:54 PM -
TY. I will do that. I assume there's no way to create a LE11 build that uses a slightly older kernel that perhaps doesn't have this bug?
I do like the idea of being able to properly test an SD card, though. Wasn't really up for figuring out how to get f3 installed on my Mac so I took the lazy route: https://github.com/vrunkel/F3XSwift
Tested my cards and they all pass.
-
I only use SanDisk "Extreme" cards after finding some knock-offs and others unreliable. 100% reliability with these cards as long as I buy them direct from SD. Hopefully they keep their quality up and not too many "clones" start to appear. For now, all good. I also have tried this on 2 other 256gb cards I have and it's entirely repeatable every time.
However, I've been wrong before, so I would be happy to test all these cards (and some other brands too), is there a "built" program anywhere for MacOS?
Ok, so moving on...
I downloaded RPI-OS 64-bit, and didn't even need to do the rpi-update, it failed without that. So, I feel pretty confident that this is a bug in the latest RPI kernel, since it doesn't show up when I install Ubuntu, Moode, LE9, LE10 or any of a number of other distros on this same card (and 2 others I have)
RPI-OS 64-bit: