Issues with 16G PNY CLass 10 SDHC card #118
Comments
|
Is it repeatable if you reformat the card and try again? How are you formatting the card? |
|
If you're installing RaspBMC using NOOBS v1.3, that explains the "big pause" during formatting on a large SD card. It's a bug #108 that's been fixed in the upcoming NOOBS v1.3.1 But that obviously doesn't explain the "No OSes in Dialog Box" part. |
Thanks Ray Doerr On Oct 31, 2013, at 3:43 AM, Rob Bishop [email protected] wrote:
|
|
What does the "some other error message" say? |
Thanks Ray Doerr On Oct 31, 2013, at 8:07 AM, maxnet [email protected] wrote:
|
|
When that happens, try:
Also if you type "dmesg" and look at the last lines, does it show SD card errors? (lines with mmcblk0 in it) |
Warning: partition 1 does not end at a cylinder boundary Then the mount of /dev/mmcblk0p1 on /mnt failed Thanks Ray Doerr On Oct 31, 2013, at 8:31 AM, maxnet [email protected] wrote:
|
|
Warnings are normal.
Does it work if you do "mount /dev/mmcblk0p1 /mnt" on the shell, or does that also give an error? Does "dmesg" indicate SD card errors or removal (e.g. mmc0: card aaa removed)?
If SD card disappears completely the NOOBS main screen may indeed stay blank and the Internet connection will not be activated. |
Thanks Ray Doerr On Nov 2, 2013, at 4:52 AM, Rob Bishop [email protected] wrote:
|
|
Were you able to respond to @maxnet 's questions? We can't help diagnose your issue without further feedback. |
Thanks Ray Doerr On Nov 2, 2013, at 8:50 AM, Rob Bishop [email protected] wrote:
|
|
Photo with a digital camera of the screen (of "dmesg") might work as well, instead of capturing a full log through serial console. Suspect the Pi/Linux doesn't get along well with your particular model SD card, or that it gets disconnected for some reason (e.g. bad pins) though. I do think the initial formatting was ok, or it would have errored out on resizing the partition earlier during installation. |
|
IMG_0659 IMG_0660 IMG_0661 IMG_0662 IMG_0663 IMG_0664 IMG_0665 IMG_0666 |
|
Try something like http://postimage.org/ |
https://picasaweb.google.com/rdoerr10/NOOBS_1_3 Thanks Ray Doerr On Nov 2, 2013, at 5:50 PM, maxnet [email protected] wrote:
|
|
Thanks for the photos. As it shows in photo 7 writing to the SD card starts giving errors (lost page write due to I/O error), and eventually after 2 minutes the connection with the card is lost completely "[ 120.771745 ] mmc0: card 0007 removed" There is a possibility that there is something wrong with the card itself. But there is also a possibility that the SD card model for some reason is not completely compatible with the Pi/Linux. |
|
I bought 4 of these 16G PHY class 10 cards and the same happens to all of the using the Raspberry Pi. They all work fine in my Mac Laptop. Sent from my iPhone5
|
|
I did as you suggested and ran the f3write and f3read on this 16G PNY CLass 10 SDHC card and it passed with no errors. Prior to running this test I used the Mac Disk Utility and erase the card and create one Fat32 partition. I think the issues NOOBS is having with these cards is related to the resizing of the partition and the creation of the extra partitions for Linux. One errors was pointing to the number of Cluster, Heads and Sectors. Thanks Ray rdoerr-mac:f3-2.2 rdoerr$ ./f3write /Volumes/UNTITLED Data OK: 14.90 GB (31256512 sectors) |
|
There is no problem with number of clusters heads etc... That's an old numbering scheme that is not used anymore. The problem could be an issue with that particular SD card model (we've seen problems like this in the past)... I'm currently working on SD card corruption as you can see from the above. We already know of problems reading from Sandisk Ultra cards and have seen one specific SD card being unable to write correctly due to a bug in the hardware. It might be worth trying out my kernel to see if the problem is the same... Will post a link to my kernel on Monday Gordon From: rdoerr01 <[email protected]mailto:[email protected]> I did as you suggested and ran the f3write and f3read on this 16G PNY CLass 10 SDHC card and it passed with no errors. Prior to running this test I used the Mac Disk Utility and erase the card and create one Fat32 partition. I think the issues NOOBS is having with these cards is related to the resizing of the partition and the creation of the extra partitions for Linux. One errors was pointing to the number of Cluster, Heads and Sectors. Thanks Ray rdoerr-mac:f3-2.2 rdoerr$ ./f3write /Volumes/UNTITLED Creating file 2.fff ... OK! Creating file 3.fff ... OK! Creating file 4.fff ... OK! Creating file 5.fff ... OK! Creating file 6.fff ... OK! Creating file 7.fff ... OK! Creating file 8.fff ... OK! Creating file 9.fff ... OK! Creating file 10.fff ... OK! Creating file 11.fff ... OK! Creating file 12.fff ... OK! Creating file 13.fff ... OK! Creating file 14.fff ... OK! Creating file 15.fff ... OK! Free space: 0.00 Byte Data OK: 14.90 GB (31256512 sectors) Reply to this email directly or view it on GitHubhttps://github.com//issues/118#issuecomment-27636432. |
|
I'm currently looking into another issue with this card, where the card with not reboot after halting. Even if you power off the Pi and remove the card... It seems you have to plug it into a debian box and mount the partitions and then unmount them to get it to work again! Sounds like the card is very 'special' |
Thanks Ray Doerr On Nov 6, 2013, at 2:19 PM, ghollingworth [email protected] wrote:
|
|
Or at least you've not noticed any problems... That doesn't mean a problem isn't waiting for just the right set of circumstances Gordon |
|
Sounds like some kind of filesystem metadata thing? Could it be that there's some kind of "unclean unmount" occurring on the Pi, which puts the first FAT filesystem into a state that the (limited) bootcode in the GPU is unable to boot from? And then Debian is able to "reset" that metadata... |
|
No it looks like Debian isn't touching the FAT partition at all... So it looks like after the halt the SD card behaves slightly differently to before meaning we're unable to read the bootcode from it The diff of the before and after inserting into the debian machine (which fixes the card) shows the first difference is right up in the ext4 partition which the bootcode doesn't read. We know that the chip is unable to even read bootcode.bin and therefore the problem must be some timing type issue with that particular type of card. I'm going to get hold of one and have a look with my SD protocol analyser, my best guess would be that the SD card holds data in some 'special' area of memory (probably SLC NAND to make it quicker) and that when you open the SDCard for writing it uses this area. When you shut down the card it should write this back into the main MLC area but it seems maybe this isn't happening and instead this only happens when you actually do a subsequent write from the debian system. So we'd need to understand what it is that we're supposed to do to this card to shut it down properly and also understand what it is that the Debian PC is doing that we're not... Gordon |
|
Ahhh, so it sounds like it might be some kind of cache flush issue inside the SD card hardware? I guess that's the kind of problem that's very hard to diagnose since the only access you have to the card is through its controller chip! (which of course may be lying to you about what it's actually doing behind the scenes) |
|
Kind of a cache issue if you think of your cache as being non-volatile! But yes I belive that's the way some of these cards work.... The real proof will be in understanding what we're doing in the protocol and what the bootcode is getting back from the card |
|
Is all the work you're doing to understand these low-level SD-card failures also helpful for upstream Linux / other SD-card-using projects in general, or are these things specific to the RPi? |
|
In general all this stuff is Raspberry Pi specific... Although the bug with the Sandisk Ultra cards would trip any system up if they were able to post accesses quick as quickly |
NOOBS is also able to write to the SD card at first. Could be that you have not run into any issues on Raspbian YET. |
|
@ghollingworth if you ever did a write-up about your SD protocol analyser stuff I bet it would make a fascinating blog post on hackaday.com ;) |
Thanks Ray Doerr On Nov 7, 2013, at 4:32 AM, ghollingworth [email protected] wrote:
|
|
@rdoerr01 Just trying to chase up on some old bugs.... :) Do you still get the same errors with the same card using NOOBS v1.3.4 (which has a much newer version of the firmware files than NOOBS v1.3) ? |
|
I just tried it again using NOOBS 1.3.4 and the same things is happening. The interesting part is when the GUI comes up where it doesn't display any OS's to install the Title of the window says Version 1.3.2 Built Nov 2, 2013. I know I used the 1.3.4 version so maybe they forget to change the Version info in the Title Bar for 1.3.4. Thanks Ray |
Hmmm, that's frustrating. Any other ideas @ghollingworth ?
Yeah, see #157 |
|
Does the same error message occur? Or has it changed? One thing to try is to format the card and make sure you give it a volume label... It seems there is a bug related to accessing some cards when there is no label! Gordon |
Thanks Ray Doerr On Feb 21, 2014, at 11:21 AM, ghollingworth [email protected] wrote:
|
|
@rdoerr01 Stupid question, but it doesn't hurt to clarify - are you using NOOBS Lite or 'full NOOBS' ? |
|
Full Sent from my iPhone5
|
|
Are you able to try #118 (comment) again and paste the results to somewhere like pastebin.com ? (you mentioned ordering a TTL serial cable so that you could collect full logs) |
|
I'm seeing this error as well on a Transcend Class 6 MicroSD card. /var/log/debug dump at https://gist.github.com/adammw/9c909ee06b244ad20748 After pressing OK, a "Error mounting settings partition" message is displayed asking if I want to reformat. If I press yes, it will go though and fail, and end up with an empty OS list. Edit: It's possible that the SD card is dodgy, as I can't even flash a Raspbian image to it and get it to work either. |
|
@adammw Perhaps it's your card-reader that's at fault, rather than the card? Worth trying a different one (preferably USB as they seem to be more reliable than internal card-readers), if you can. |
|
@lurch is this still in progress? |
|
Can any of the previous commenters in this thread try re-testing with NOOBS v1.4 or above? |
|
This SD card bug is still there, this issue is duplicated somewhere. But I think its gone for Pi1 while it reappeared with the pi2 and is still there. |
|
@NicoHood The Pi2 SD-card bug is #230, which AUIU is an entirely separate bug from the one being discussed here. But with so many different SD-card related bugs (see also #265, #258, #174) it's difficult to separate them into different categories (especially when they're so difficult to try and reproduce!). |
|
Closing since OP hasn't responded since 2014 and there have been several SD card issue fixes added since then. |
When trying to install NOOBS v1.3 with the Local and Network file, it boots up fine and sits for a very long time at the step "Formatting settings partition". If you wait about 3 minutes it will finally get past that and have some other error message and then it comes up but doesn't show any OS's in the Dialog Box.
The text was updated successfully, but these errors were encountered: