gsh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 1421
Joined: Sat Sep 10, 2011 11:43 am

Re: SD Card Corruption

Tue Jun 04, 2013 9:04 pm

So if you have a card / pi combination that exhibits the problem straight away then can you please PM me your details and I'll get a new Pi / card sent and you can send me yours to test?

How reproducible is the corruption? Are we always talking many hours / days or do you have something that is more susceptible to corruption?

Gordon
--
Gordon Hollingworth PhD
Raspberry Pi - Director of Software Engineering

gkreidl
Posts: 5998
Joined: Thu Jan 26, 2012 1:07 pm
Location: Germany

Re: SD Card Corruption

Tue Jun 04, 2013 10:00 pm

Meanwhile I have to add something to my report above:

I have enabled the technisat dvbs HD usb driver again in spite of overclocking with 900 Mhz and I have had no more SD card issues for weeks. But I changed one thing. In config.txt I set:

force_turbo=1

Maybe some of the problems result from dynamic overclocking.
Minimal Kiosk Browser (kweb)
Slim, fast webkit browser with support for audio+video+playlists+youtube+pdf+download
Optional fullscreen kiosk mode and command interface for embedded applications
Includes omxplayerGUI, an X front end for omxplayer

hwintjen
Posts: 5
Joined: Thu Jul 05, 2012 5:55 am

Re: SD Card Corruption

Wed Jun 12, 2013 8:22 pm

Hi All. I have had several SD corruptions in the past also.

I own 3 RPIs and tried many SD cards.

My current impression is that it is more related to the RPI itself that to the power supply / SD card,

since i got a SD Card 16GB Sandisk worked in one RPI (512MB) for months without any SD card trouble (according to dmesg) plugged in one of my other RPIs (with which i had trouble previously already and tried some other SD cards), and that RPI smashed the filesystem in less than 2 minutes.

A coulleage of mine had bought 10 RPis under constant data write tests and confirmed that he has one RPI which makes trouble always regardless of the SD cards (all identical power supply). He had swapped the SD cards and the problems seems to manifest on one RPI.

Regards

Hauke

simplesi
Posts: 2327
Joined: Fri Feb 24, 2012 6:19 pm
Location: Euxton, Lancashire, UK
Contact: Website

Re: SD Card Corruption

Thu Jun 13, 2013 5:59 am

nd that RPI smashed the filesystem in l ... 2 minutes.
I'd take up gsh's offer and swap that one out with him so that they can take a look at it :)

I'm currently running a some cheap generic 4GB Class 4 cards and last week moved them up all up to 800MHz - after tonights Robot club, I'll try upping them a bit more

Simon
Seeking help with Scratch and I/O stuff for Primary age children
http://cymplecy.wordpress.com/ @cymplecy on twitter

rlunaro
Posts: 1
Joined: Mon Sep 09, 2013 11:45 am

Re: SD Card Corruption

Mon Sep 09, 2013 11:51 am

Hello everybody:

I humbly think that the problem comes from the USB chargers we are using. Definitely, they are not power sources. After a couple of spoiled SD cards, there is a log of my research:

http://www.supermanhamuerto.com/doku.ph ... corruption

The chargers behave strangely: they emit like "pulses" of power that definitely aren't the best for a computer. Moreover, most of them have "5V" in their specifications and this is not true. Multimeter in hand, most of them are 7.5V, two miles far away from the desired 5V.

At the present moment I am trying to grab a proper power source and see what happens.

More things: in this thread I can read that many recent bulks of raspberries haven't showed this problem. This can be due to a change in the desing of the board, the range of specifications or something as simple as many raspberries nowadays are sold with a "pack" which includes the case and the power source.

DanR
Posts: 33
Joined: Fri Jan 18, 2013 1:20 pm

Re: SD Card Corruption

Wed Feb 05, 2014 8:49 pm

I've found that one of my RPi's which has the self resettable fuses on the usb power lines causes me issues but the other RPi with 0ohm links instead of the fuses works perfectly with the same card type as the one that failed. Both were overclocked.

I've soldered 2010 style 0ohm resistors over the top of the fuses and I'll test to see if that helps cure the issue. I'll report back just in case this helps someone else.

DanR
Posts: 33
Joined: Fri Jan 18, 2013 1:20 pm

Re: SD Card Corruption

Thu Feb 06, 2014 4:18 am

Well that didn't work either, the card fails after some time on both Rpi's. I'm going to pursue this as it's driving me nuts! I hope the parallella board (when it arrives) doesn't exhibit this issue :shock:

DaveyWavey
Posts: 6
Joined: Sat Aug 30, 2014 8:45 am

Re: SD Card Corruption

Sat Aug 30, 2014 9:39 am

(I know this is a slightly old thread, but I felt that it was useful to share my experiences, and it didn't seem helpful to start a whole new thread on the same topic. Sorry if that irks some people.)

A couple of weeks ago, I was also experiencing the same symptoms as many others on my Raspberry Pi B+, running Raspbian (Linux raspberrypi 3.12.26+ #704 PREEMPT) from a Sandisk 64 GB microSDHC card.

It has been working just fine for a week or two, but then started becoming unresponsive, requiring a power cycle. I wasn't seeing anything interesting in the system logs, so configured rsyslog to forward log messages to a remote syslog server (in case the Pi wasn't able to write pertinent errors to the SD card).

Long story short, the next time this happened I was able to spot the problem in the log messages that were forwarded to my syslog server. The smoking gun appeared to be a sequence of I/O errors, followed by a "aborting journal on device" message, then a "remounting filesystem read-only". Obviously, based on the symptoms (and a load of similar experiences shared on these forums), I assumed this indicated some sort of SD card error. The problems always seemed to coincide with periods of increased I/O activity (e.g. downloading large number of new packages via apt-get).

So, I decided to re-format the SD card and re-install everything from scratch. I reformatted it using the official SD formatter (https://www.sdcard.org/downloads/formatter_4/), then re-wrote the Rasbian image to it, and started again. However it failed again quite quickly (same types of I/O and filesystem errors) during an apt-get upgrade operation.

So I concluded that my SD card was knackered and bought a new one (this time a Samsung 16 GB card, listed here as 'working': http://elinux.org/RPi_SD_cards) and started again with this one. To my intense irritation, the problem recurred again really soon with this card.

I decided that my least bad option would be to try to reduce write operations on my SD card as much as possible, so I bought a 64GB SSD drive (http://www.sandisk.co.uk/products/ssd/sata/standard/), reverted to my original SD card (after all, it had previously worked OK for at least a week, unlike the other one), re-installed Raspbian, then followed paulv's excellent instructions (http://www.raspberrypi.org/forums/viewt ... 29&t=44177) to move the root filesystem to the SSD drive -- effectively just using the SD card at boot-time, but then running the OS from the SSD drive via USB. My hope was that this would minimise use of the SD card, whilst using a better-suited drive to so the vast majority of the I/O.

Imagine my annoyance, then, when after 30 minutes of running just fine from the SSD drive, the same problems occurred again (except this time the errors pointed at the filesystem on my SSD drive, not the boot partition on my SD card). At this point, I was finding it hard to believe that I could have somehow chosen two different SD cards and one SSD drive all with the same defects and/or incompatibilities. I now had to assume that this was a hardware problem with the Pi itself.

So, I replaced the Pi with another B+ that I had lying around. I changed nothing else -- same SD card, same SSD drive, same power supply, etc. Everything worked fine, and it's been working flawlessly ever since.

I have to conclude then, that although my case initially showed all the symptoms of a problematic/corrupted SD card, in fact the problem was with the Pi itself. If I hadn't taken the unusual step of switching to an SSD drive, I would have probably kept on assuming that these were just SD card problems.

It's hard to imagine what the true cause of the problem with the Pi were -- evidently something that affected I/O via the SD card interface and USB, but I don't know enough about the Pi's hardware design to speculate about what that could be.

Anyway, apologies for the long rambling story, however I thought it would be useful to share this, in case it sheds any light on other people's problems. While I'm sure that many suspected SD card problems are just that: problems with the SD card, it's worth bearing in mind that in at least one case, the problem was nothing to do with the SD card, and lay with the Pi itself.

RobbyB
Posts: 1
Joined: Thu Sep 18, 2014 11:59 am

Re: SD Card Corruption

Thu Sep 18, 2014 12:28 pm

I just had a case of SDcard corruption and my solution was rather simple: After cleaning the contacts on the card with a glass brush the card is working again !

ArashJaberi
Posts: 1
Joined: Thu Sep 18, 2014 1:26 pm

Re: SD Card Corruption

Thu Sep 18, 2014 1:29 pm

I think bonnie++ is a great test to detect bad sectors as well.

amzi2000
Posts: 1
Joined: Tue Oct 14, 2014 12:01 pm

Re: SD Card Corruption

Tue Oct 14, 2014 12:49 pm

Hi,

Managed to get SD card corruption also. Pi suddenly just crashed on me after working fine for about six weeks.

Now boot fails in two separate Raspberries in the same spot:
"
EXT4-fs .... Checksum for group 0 failed ....
EST4-fs ... group descriptors corrupted!
...
No filesystem could mount root, tried ext4
Kernel panic, etc
"

This seems to be a known issue - does anybody have ideas how to fix or go around it? I can format the card and reinstall but how to prevent this from happening again?

kritik
Posts: 1
Joined: Sun Nov 02, 2014 11:34 am

Re: SD Card Corruption

Sun Nov 02, 2014 11:45 am

Faced the same problem. No overclocking, just distro/packets installation. Ext filesystem check says that it's ok. Then I tried to clean elements with vodka and still the same problem.Just to mention electricity at RPi place is very stable and card producer is Samsung.

Return to “Troubleshooting”