BBB flasher fails to update eMMC

Enter code here...

Hi All,

I have been trying to update the eMMC on a BBB that I just received with no success. The board booted off the eMMC when I first received it.

Following the steps at http://beagleboard.org/Getting%20Started I got to the step “Update board with latest software” and downloaded the Angstrom version dated 2013-06-20 from http://beagleboard.org/latest-images by following the links on the getting started page. Later I discovered there is a newer version of the flasher image at http://circuitco.com/support/index.php?title=Updating_The_Software called the Production Image 2013.09.04 and downloaded it.

I had previously installed the utilities needed, and successfully expanded and copied the image to a 4G SD card. I booted from the SD card using a 1A power supply and everything seemed to work as expected for about 5 minutes, then all the user LEDs turned off and nothing more happened.

At this point the board would no longer boot from the eMMC, but I could boot from the SD card. I downloaded and burned the image called SD Boot 2013.08.21 (the current production image) from the circuitco.com Updating the Software page to a different SD card. The BBB boots from that SD card and runs as expected. I reloaded the first SD card with that SD Boot image to verify that the SD card worked correctly, and it did. I then reloaded it with the flasher image and tried again. Same result, all LEDs off and appeared hung after about 5 minutes. I had read that the eMMC flashing process takes about 45 minutes, so I tried waiting for an hour and the LEDS never changed. The board still would not boot from the eMMC after the long wait.

I connected a serial debug cable to see if there was an error that would explain what might be going wrong, but I didn’t find anything useful. However, the serial cable did allow me to login and poke around after the LEDs turned off. I found the page at http://www.crashcourse.ca/wiki/index.php/BBB_software_update_process which explains how the flasher image is supposed to work and I dug into the emmc.sh and mkcard.sh scripts that do the flashing.

Seeing nothing out of the ordinary I tried running the emmc.sh script manually so I could see the progress messages and any error messages. Those messages are copied below.

`

root@beaglebone:~# /usr/bin/emmc.sh 2>&1 | tee emmc.log
Valid EEPROM header found
/usr/bin/emmc.sh: line 35: blockdev: command not found
Paritioning eMMC
16+0 records in
16+0 records out
1+0 records in
1+0 records out
DISK SIZE - 1920991232 bytes
CYLINDERS - 233
Checking that no-one is using this disk right now …
OK

Disk /dev/mmcblk1: 233 cylinders, 255 heads, 63 sectors/track

sfdisk: ERROR: sector 3069161472 does not have an msdos signature
sfdisk: /dev/mmcblk1: unrecognized partition table type

sfdisk: No partitions found

Old situation:
New situation:
Units = cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0

Device Boot Start End #cyls #blocks Id System
/dev/mmcblk1p1 * 0+ 8 9- 72261 c W95 FAT32 (LBA)
/dev/mmcblk1p2 9 232 224 1799280 83 Linux
/dev/mmcblk1p3 0 - 0 0 0 Empty
/dev/mmcblk1p4 0 - 0 0 0 Empty
Successfully wrote the new partition table

Re-reading the partition table …

If you created or changed a DOS partition, /dev/foo7, say, then use dd(1)
to zero the first 512 bytes: dd if=/dev/zero of=/dev/foo7 bs=512 count=1
(See fdisk(8).)
./mkcard.sh: line 38: kpartx: command not found
umount: /dev/mmcblk1p1: not mounted
mkfs.vfat 2.11 (12 Mar 2005)
umount: /dev/mmcblk1p2: not mounted
mke2fs 1.42.1 (17-Feb-2012)
Discarding device blocks: done
Filesystem label=Angstrom
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
Stride=0 blocks, Stripe width=0 blocks
112672 inodes, 449820 blocks
22491 blocks (5.00%) reserved for the super user
First data block=0
Maximum filesystem blocks=461373440
14 block groups
32768 blocks per group, 32768 fragments per group
8048 inodes per group
Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912

Allocating group tables: done
Writing inode tables: done
Creating journal (8192 blocks): done
Writing superblocks and filesystem accounting information: done

./mkcard.sh: line 82: kpartx: command not found
Mounting partitions
Copying bootloader files
/usr/bin/emmc.sh: line 54: blockdev: command not found
Extracting rootfs
Populating VFAT partition
/usr/bin/emmc.sh: line 66: /media/2/etc/dogtag: No such file or directory
Generating machine ID
Spawning namespace container on /media/2 (console is /dev/pts/0).
Timezone Universal does not exist in container, not updating container timezone.
execv() failed: No such file or directory
Container failed with error code 1.
Running Postinsts
Spawning namespace container on /media/2 (console is /dev/pts/0).
Timezone Universal does not exist in container, not updating container timezone.
execv() failed: No such file or directory
Container failed with error code 1.
Spawning namespace container on /media/2 (console is /dev/pts/0).
Timezone Universal does not exist in container, not updating container timezone.
Container failed with error code 1.
ln: /media/2/etc/systemd/system/xinetd.service: No such file or directory
touch: /media/2/etc/default/locale: No such file or directory
/usr/bin/emmc.sh: line 111: blockdev: command not found
ID.txt found
START.htm missing - ERROR
md5sum: can’t open ‘/media/2/usr/share/pixmaps/backgrounds/gnome/angstrom-default.jpg’: No such file or directory
Wallpaper MD5sum failed
/usr/bin/emmc.sh: line 153: blockdev: command not found
100000+0 records in
100000+0 records out
ERRORS found: , START.htm, bgmd5
root@beaglebone:~#

`

There are a couple of problems with these scripts. First, the mkcard.sh script uses a bad test to see if the kpartx command exists. The test used always passes and the the script tries to execute kpartx even though it is not installed in the flasher image. This leads to the following errors:

`

./mkcard.sh: line 38: kpartx: command not found ./mkcard.sh: line 82: kpartx: command not found

`

I think the script should quote the output of the which command before testing if it is executable like this. Without the double quotes the test is always true if the which command returns nothing (i.e. when the kpartx commnad doesn’t exist).

`

if [ -x “which kpartx” ]; then
kpartx -a -v ${DRIVE}
fi

`

Second the emmc.sh script tries to execute the blockdev command a few times leading to the following errors because this command is not available in the flasher image.

`

/usr/bin/emmc.sh: line 35: blockdev: command not found /usr/bin/emmc.sh: line 54: blockdev: command not found /usr/bin/emmc.sh: line 111: blockdev: command not found /usr/bin/emmc.sh: line 153: blockdev: command not found

`

After the “Extracting rootfs” progress message there are several errors saying no such file or directory. The first is after the script tries to write to a file.

`
echo “Cloud9 GNOME Image 2013.09.04” > ${PART2MOUNT}/etc/dogtag

`

The problem is that the /etc directory does not exist in the filesystem created on the second partition. This filesystem is created by expanding the tar archive into the linux partition on the eMMC.

`

echo “Extracting rootfs”
tar zxf Angstrom-Cloud9-IDE-GNOME-eglibc-ipk-v2012.12-beaglebone.rootfs.tar.gz -C ${PART2MOUNT}

`

Using the -t option to tar to list the contents rather than expanding them I looked at the contents of the rootfs archive.

`

./
./bin/
./bin/egrep
./bin/chmod
./bin/chmod.coreutils
./bin/chgrp
./bin/ps.procps
./bin/traceroute6
./bin/chattr
./bin/ls.coreutils
./bin/deluser
./bin/mount.util-linux
./bin/systemd-notify
./bin/more.util-linux
./bin/pidof
./bin/dnsdomainname
./bin/cpio
./bin/dmesg.util-linux
./bin/tar
./bin/pwd.coreutils
./bin/adduser
./bin/sed.sed
./bin/kill.util-linux
./bin/dd
./bin/df
./bin/busybox
./bin/date.coreutils
./bin/date
./bin/stty
./bin/grep.grep
./bin/addgroup
./bin/ps
./bin/rmdir.coreutils
./bin/zcat
./bin/uname.coreutils
./bin/tracepath6
./bin/ping.iputils
./bin/dmesg
./bin/tinylogin
./bin/dd.coreutils
./bin/cp
./bin/rm.coreutils
./bin/gunzip
./bin/ls
./bin/pidof.sysvinit
./bin/netstat
./bin/tempfile.debianutils
./bin/true.coreutils
./bin/false.coreutils
./bin/ping
./bin/mkdir.coreutils
./bin/ping6
./bin/gzip
./bin/mkdir
./bin/mknod
./bin/more
./bin/rm
./bin/tar.tar
./bin/vi
./bin/run-parts.debianutils
./bin/ash
./bin/stty.coreutils
./bin/sleep.coreutils
./bin/ip
./bin/systemd-inhibit
./bin/ln.coreutils
./bin/sh
./bin/cat.coreutils
./bin/systemd-tmpfiles
./bin/tracepath
./bin/dumpkmap
./bin/su
./bin/systemd-machine-id-setup
./bin/login
./bin/umount.util-linux
./bin/sync.coreutils
./bin/systemd-tty-ask-password-agent
./bin/mount
./bin/gzip.gzip
./bin/mknod.coreutils
./bin/cat
./bin/sed
./bin/chgrp.coreutils
./bin/echo.coreutils
./bin/zcat.gzip
./bin/cp.coreutils
./bin/touch
./bin/ln
./bin/arping
./bin/chown.coreutils
./bin/touch.coreutils
./bin/bash
./bin/hostname
./bin/lsmod
./bin/kmod
./bin/systemd-ask-password
./bin/base64
./bin/rmdir
./bin/sleep
./bin/fgrep.grep
./bin/lsmod.kmod
./bin/sync
./bin/umount
./bin/usleep
./bin/uname
./bin/loginctl
./bin/delgroup
./bin/chown
./bin/fgrep
./bin/kill.procps
./bin/gunzip.gzip
./bin/grep
./bin/login.shadow
./bin/ping6.iputils
./bin/echo
./bin/systemctl
./bin/kill.coreutils
./bin/mktemp
./bin/true
./bin/false
./bin/journalctl
./bin/pwd
./bin/mv
./bin/mv.coreutils
./bin/egrep.grep
./bin/kill
./sbin/
./sbin/rtmon
./sbin/modprobe
./sbin/switch_root
./sbin/fdisk
./sbin/blockdev.util-linux
./sbin/hdparm
./sbin/insmod.kmod
./sbin/routel
./sbin/ifconfig
./sbin/pivot_root.util-linux
./sbin/swapon
./sbin/logread
./sbin/fsck.minix.util-linux
./sbin/mkfs.minix
./sbin/killall5
./sbin/setconsole
./sbin/e4defrag
./sbin/fsck.minix
./sbin/vipw
./sbin/losetup.util-linux
./sbin/logsave
./sbin/zcav
./sbin/mklost+found
./sbin/fsck.ext4
./sbin/debugfs
./sbin/sysctl.procps
./sbin/reboot
./sbin/route
./sbin/getty.tinylogin
./sbin/ctrlaltdel
./sbin/hdparm.hdparm
./sbin/insmod
./sbin/routef
./sbin/losetup
./sbin/mkfs.ext4dev
./sbin/depmod.kmod
./sbin/ifcfg
./sbin/modprobe.kmod
./sbin/rmmod.kmod
./sbin/getty
./sbin/dumpe2fs
./sbin/readprofile
./sbin/sysctl
./sbin/halt
./sbin/depmod
./sbin/mkswap.util-linux
./sbin/loadkmap
./sbin/fsck.ext3
./sbin/e2fsck
./sbin/ldconfig
./sbin/udhcpc
./sbin/mkfs.ext2
./sbin/mkfs.ext4
./sbin/bonnie++
./sbin/umount.udisks
./sbin/vigr
./sbin/ifup
./sbin/fsck.ext4dev
./sbin/rtpr
./sbin/resize2fs
./sbin/hwclock
./sbin/swapoff.util-linux
./sbin/lsmod
./sbin/modinfo
./sbin/start-stop-daemon
./sbin/pivot_root
./sbin/blockdev
./sbin/init
./sbin/vigr.shadow
./sbin/agetty
./sbin/nologin
./sbin/mkfs.minix.util-linux
./sbin/sulogin
./sbin/fsck
./sbin/badblocks
./sbin/ip.iproute2
./sbin/swapon.util-linux
./sbin/cfdisk
./sbin/e2image
./sbin/fsck.ext2
./sbin/klogd
./sbin/tc
./sbin/vipw.shadow
./sbin/syslogd
./sbin/filefrag
./sbin/rmmod
./sbin/fsck.util-linux
./sbin/hwclock.util-linux
./sbin/poweroff
./sbin/swapoff
./sbin/shutdown
./sbin/mkswap
./sbin/e2undo
./sbin/e2freefrag
./sbin/mkfs.ext3
./sbin/modinfo.kmod
./sbin/mke2fs
./sbin/fdisk.util-linux
./sbin/ifdown
./media/
./media/card/
./media/realroot/
./media/mmc1/
./media/ram/
./media/net/
./media/hdd/
./media/cf/
./media/union/
./home/
./home/xuser/
./home/root/
./mnt/
./mnt/card
./mnt/ram
./mnt/net
./mnt/cf
./usr/
./usr/bin/
./usr/bin/exo-desktop-item-edit
./usr/bin/identify
./usr/bin/realpath
./usr/bin/colrm
./usr/bin/sensible-browser.debianutils
./usr/bin/dvbv5-scan
./usr/bin/wvRTF
./usr/bin/modutil
./usr/bin/aconnect
./usr/bin/cytune
./usr/bin/sg_write_same
./usr/bin/timeout
./usr/bin/strsclnt
./usr/bin/xargs
./usr/bin/eu-ar
./usr/bin/pkaction
./usr/bin/wvText
./usr/bin/prlimit
./usr/bin/wvPS
./usr/bin/arm-angstrom-linux-gnueabi-nm
./usr/bin/tset
./usr/bin/vdir.coreutils
./usr/bin/sensible-pager.debianutils
./usr/bin/sum.coreutils
./usr/bin/gpg-error-config
./usr/bin/sg_sat_set_features
./usr/bin/expr
./usr/bin/x11vnc
./usr/bin/wvDVI
./usr/bin/resolv-test
./usr/bin/dbus-daemon
./usr/bin/ivtv-ctl
./usr/bin/rsaperf
./usr/bin/gst-xmlinspect-0.10
./usr/bin/arecord
./usr/bin/roff2pdf
./usr/bin/vim
./usr/bin/lesskey
./usr/bin/sg_luns
./usr/bin/free
./usr/bin/newgrp
./usr/bin/afmtodit
./usr/bin/users.coreutils
./usr/bin/bzcat
./usr/bin/sg_vpd
./usr/bin/refer
./usr/bin/sg_write_buffer
./usr/bin/gnome-session
./usr/bin/gvfs-set-attribute
./usr/bin/freetype-config
./usr/bin/sftp
./usr/bin/md5sum
./usr/bin/arm-angstrom-linux-gnueabi-g++
./usr/bin/private-network-test
./usr/bin/wvWare
./usr/bin/pk-example-frobnicate
./usr/bin/eu-elfcmp
./usr/bin/run-parts
./usr/bin/evtest-capture
./usr/bin/dvb-fe-tool
./usr/bin/xfce4-about
./usr/bin/avahi-resolve
./usr/bin/tsort
./usr/bin/mcookie
./usr/bin/compile_et
./usr/bin/csplit
./usr/bin/strip
./usr/bin/clear
./usr/bin/arm-angstrom-linux-gnueabi-readelf
./usr/bin/lslocks
./usr/bin/localectl
./usr/bin/xzcmp
./usr/bin/pip_stress
./usr/bin/ionice
./usr/bin/setarch
./usr/bin/chattr
./usr/bin/pre-grohtml
./usr/bin/xrandr
./usr/bin/avconv
./usr/bin/pwdx.procps
./usr/bin/rnano
./usr/bin/xmlcatalog
./usr/bin/expiry
./usr/bin/xkbcomp
./usr/bin/sg_ses
./usr/bin/sg_rtpg
./usr/bin/dirname.coreutils
./usr/bin/gnome-keyring-daemon
./usr/bin/bzless
./usr/bin/xzfgrep
./usr/bin/arm-angstrom-linux-gnueabi-as
./usr/bin/perl5.14.2
./usr/bin/sha384sum
./usr/bin/chsh.shadow
./usr/bin/basename.coreutils
./usr/bin/desktop-file-validate
./usr/bin/xdg-desktop-menu
./usr/bin/pathchk
./usr/bin/sg_unmap
./usr/bin/eqn2graph
./usr/bin/script
./usr/bin/g++
./usr/bin/savelog.debianutils
./usr/bin/gtester-report
./usr/bin/hcitool
./usr/bin/sg_prevent
./usr/bin/arm-angstrom-linux-gnueabi-cpp
./usr/bin/scp
./usr/bin/p7content
./usr/bin/gvfs-monitor-dir
./usr/bin/pmap.procps
./usr/bin/flushdisk
./usr/bin/vbi-test
./usr/bin/dbclient
./usr/bin/bluetooth-properties
./usr/bin/grn
./usr/bin/fuser
./usr/bin/btoa
./usr/bin/tftp
./usr/bin/xminicom
./usr/bin/chcon
./usr/bin/eu-make-debug-archive
./usr/bin/runcon.coreutils
./usr/bin/sha1sum.coreutils
./usr/bin/flex
./usr/bin/metacity-message
./usr/bin/update-alternatives
./usr/bin/gvfs-tree
./usr/bin/gnome-desktop-item-edit
./usr/bin/pp
./usr/bin/wvCleanLatex
./usr/bin/uptime
./usr/bin/tlb
./usr/bin/cc
./usr/bin/eject
./usr/bin/certcgi

`

As you can see there is no /etc directory in this archive, hence no /etc directory in the filesystem created on the eMMC linux partition. It seems to me this archive is missing some key pieces.

Later, near the end of the messages, the script reports that START.htm can’t be found.

`
START.htm missing - ERROR

`

This file is supposed to be copied from the directory /usr/share/beaglebone-getting-started in the new linux partition, but that directory is also missing from the archive.

Next, the wallpaper checksum test fails because the directory /usr/share does not exist. This directory is also missing from the archive.

`
md5sum: can’t open ‘/media/2/usr/share/pixmaps/backgrounds/gnome/angstrom-default.jpg’: No such file or directory
Wallpaper MD5sum failed

`

Finally, the user LEDs are turned off by the script because of the previous errors by the test at the end of the emmc.sh script.

`

echo “ERRORS found: ${ERROR}”
if [ -e /sys/class/leds/beaglebone:green:usr0/trigger ] ; then
echo none > /sys/class/leds/beaglebone:green:usr0/trigger
echo none > /sys/class/leds/beaglebone:green:usr1/trigger
echo none > /sys/class/leds/beaglebone:green:usr2/trigger
echo none > /sys/class/leds/beaglebone:green:usr3/trigger
fi

`

So now I know why the LEDs are off (because the flasher script turned them off) and why it takes about 5 minutes to fail (because the rootfs archive in the flasher image is missing key parts, so it is expanded in 5 minutes instead of 45 minutes).

Perhaps the flasher image I downloaded is corrupt, but I have no easy way to verify that because there are no MD5 sums posted for the image files along with the download links. I suspect the image is OK but I am downloading it again to compare the two files.

I know this is a long post, but I thought the detail might help someone to fix these errors and make the flashing process run more smoothly, since I see that many people have had problems when doing this. Also, I suspect that this update may have been a waste of my time since I think my BBB may have had this 09.04 version of the software installed in the eMMC when it shipped last week. It would be helpful to add the info about checking the current version from the top of the http://circuitco.com/support/index.php?title=Updating_The_Software page to the Getting Started page so that users can check if they really need to do an update before they waste a lot of time.

I look forward to working with my new BBB.

Dennis Cote

I will add it to the top of the page for those that only read part of the page. I just hope they start at the top.

Gerald

Hi Gerald,

Just to be clear, I think the new info should be added in the middle of http://beagleboard.org/Getting%20Started between the “Update board with latest software” and the “Step #1: Download the latest software image”. It should explain how to check the current version so the user can skip the download and update if they already have the current version (which is likely on a new board).

I have downloaded the 2013.09.04 flasher image again and verified that I get the same MD5 hash as the original file I downloaded, so I’m sure the download was not corrupted. Do you have any idea when a new flasher image with a complete rootfs archive will be available?

Dennis Cote

Just to be clear, I agree. But, I handle the Wiki portion. I will pass that suggestion on to someone else to see if that can be updated. We point people to the WIKI via the card that comes in the box as that is easier t update than the website…

I will need to defer your question on the image to the rest of the team to answer.

Gerald

I will pass that suggestion on to someone else to see if that can be updated.

Thanks.

We point people to the WIKI via the card that comes in the box as that is easier t update than the website…

The card is what led me to the circuit.co wiki page with the 2013.09.04 image. I had already downloaded the older 2013-06-20 image linked to from the Getting Started web page. Maybe the web page can be updated with a link to the wiki labeled “for the latest info go to…” rather than relying on just the cards.

I will need to defer your question on the image to the rest of the team to answer.

OK, hopefully someone can help. I don’t see how anyone could have used this image successfully, but I may be wrong.

Has anyone used the 2013.09.04 flasher image to install the latest Angstrom version on a BBB eMMC?

Dennis Cote

That is the BeagleBoard.org Wiki. Circuitco is just the current host. That should get changed soon to another host location.

We use the 9_4 Flasher image in production to flash all the production boards. It is tested thousands of time a week.

Gerald

let me share my experience (problems).

  1. the card can be 4gb but the use of the card it must be the sice of the bbb memory 2gb

  2. when you restore is not ok to use any thing else (usb share, ethernet, linux session). just dont do any thing just press the boot bottom.

  3. try more than one time. and be patient to wait.

  4. before you use the image to the card, the arquive that you download is compress. decompress it first and than copie to the sd card . you can only copy the .img file.

that image is very good. I use ubuntu-precise-12.04.2-armhf-3.8.13-bone20.img and Angstrom-Cloud9-IDE-GNOME-eglibc-ipk-v2012.12-beaglebone-2013.06.20.

Can you verify that the file linked to from the wiki at https://s3.amazonaws.com/angstrom/demo/beaglebone/BBB-eMMC-flasher-2013.09.04.img.xz is exactly the same as the file used to build the production flasher SD cards?

The file I downloaded is BBB-eMMC-flasher-2013.09.04.img.xz and it is 373,850,348 bytes long with MD5 hash 6b551bf357339e98377dfb6f6fdba016.

It is expanded by 7zip into an image file BBB-eMMC-flasher-2013.09.04.img which is 3,657,433,088 bytes long and has MD5 hash b45fcc0da3d797f6dc0fed3dde19d311.

This image is copied (using the Win32DiskImager) to a 4GB SD card which boots the BBB and runs the emmc.sh script as expected, but the script consistently fails after about 5 minutes as explained in my previous post. The board stops with all user LEDs off because of errors detected by the emmc.sh script. Those errors are caused by missing files and directories in the archived rootfs, Angstrom-Cloud9-IDE-GNOME-eglibc-ipk-v2012.12-beaglebone.rootfs.tar.gz, contained in the /build directory of the Angstrom linux installed on the flasher SD card.

I just completed another pass through the entire update process starting with the freshly downloaded compressed flasher image. It failed in the same way. After the emmc.sh script stopped I logged in using the serial debug cable and tried to expand the archived rootfs manually. I got an error about an invalid tar header checksum.

`

root@beaglebone:~# cd /build
root@beaglebone:/build# mkdir rootfs
root@beaglebone:/build# tar -zxf Angstrom-Cloud9-IDE-GNOME-eglibc-ipk-v2012.12-beaglebone.rootfs.tar.gz -C rootfs
tar: invalid tar header checksum
root@beaglebone:/build# ls rootfs
bin home media mnt sbin usr

`

The directories created look the same as those listed in my previous post. I just repeated the tar listing again and go the same error. I must have missed it last time since I had redirected the output to a file, but the error message went to the console instead of the file.

There is definitely a problem with the flasher image I downloaded. The archived rootfs is corrupt so it generates an incomplete rootfs in the eMMC.

Dennis Cote

That is the one we are using. But I will have it double checked.

Gerald

... snip ...

  good lord, i step out for a few hours and there's flour and cat
prints all over the kitchen when i get back.

  given that i've done *loads* of reflashes of my BBBs, i can test
this -- i'll grab the file

http://dominion.thruhere.net/koen/angstrom/beaglebone/BBB-eMMC-flasher-2013.09.04.img.xz

and check it out, is that the file that is currently the subject of
debate?

rday

No. The link above that:

https://s3.amazonaws.com/angstrom/demo/beaglebone/BBB-eMMC-flasher-2013.09.04.img.xz

Gerald

The file I downloaded is BBB-eMMC-flasher-2013.09.04.img.xz and it
is�373,850,348 bytes long with MD5
hash�6b551bf357339e98377dfb6f6fdba016.

  verified with mine ...

It is expanded by 7zip into an image
file�BBB-eMMC-flasher-2013.09.04.img which is�3,657,433,088 bytes
long and has MD5 hash�b45fcc0da3d797f6dc0fed3dde19d311.

  verified with mine, it's now reflashing one of my BBBs, it's been
cranking away for about 8 minutes now and still going ...

rday

which is *precisely* why i chose a different one ... so that if mine
works, we have at least isolated the problem to a particular tarball.
10 minutes in, and still flashing ...

rday

So if yours works then the S3 version is definitely bad. Sound logic.

Gerald

still chugging away so i'm confident it will finish. i did nothing
different from all the other times:

* download xz-ball
* uncompress to get .img file
* verified that my size and md5sum matched what original poster listed
* "dd" to 4G uSD card
* remove uSD card, then immediately reinsert to verify that
  filesystems are mounted by my linux box
* put uSD card in BBB and ... go

  the fact that my size and md5sum matched suggests something gets
corrupted in the flashing process or something. i've always downloaded
my flasher images from koen's dominion site, never from amazon.

rday

One of the things that will affect flashing is if say Ethernet is connected. That is why I say do not connect it. Anything on the USB host port such as a terminal could I suppose also affect it.

Another thing is trying to flash while using USB power.

Gerald

Hi Robert,

I agree that if your MD5 checksums matches mine the problem must be in the writing of the image to the SD card. I am using Win7 instead of linux to expand the image and write the SD card, but I don’t think that should matter. I will try writing the SD card with a different SD card adapter on a different PC.

BTW I can’t get the http://dominion.thruhere.net/koen/angstrom/beaglebone/BBB-eMMC-flasher-2013.09.04.img.xz image since the domain is blocked by our untangle firewall for the following reason.

`
This site was categorized in: Parked Domains

`

I will have to talk to our network admin about this, but I believe we use the default blacklists for untangle. Since others may use this firewall (http://www.untangle.com/store/firewall.html) you may want to see if they can correct their blacklist.

Dennis Cote

I use Win7 as well. May be an issue with the unzip program you are using.

Gerald

flashing worked fine for me. my recipe (aside from the earlier
stuff) is, as you say:

* 5V power
* nothing else connected

the only weird issue i had once upon a time is when i had one of
specialcomp's USB-TTL adapters connected to the 6-pin header -- man,
the boot process didn't like that.

  but, as i said, i had no problem flashing here.

rday

i can't think of anything else ... if my size and md5sum matches
yours after decompression, there's really no other possibility.

rday