Unable to update my machine

classic Classic list List threaded Threaded
11 messages Options
Reply | Threaded
Open this post in threaded view
|

Unable to update my machine

crazy_bus
To Devel list

I've got a b4 which I've upgraded once when I first got it and am unable to
update since.  I'm using the same usb flash key as I used the first time.  I
first tried it unsuccessfully with all the files the same except for updated
firmware .  Since then I've tried multiple times with the same files and
later with the newer firmware and newer build.  I've also tried re
downloading the files along with reformatting the stick (as both FAT 16 and
32)to see if it helps.

When I insert the stick and turn on the machine with the x key pressed it
recognises the key is there but doesn't seem to see the script.  As it
doesn't  mention that it's either upgrading the machine or if it's firmware
is current.
The only difference I see after this happening is the usb stick now has a
folder on it called .olpc.store

The usb stick has only one partition and as far as I know sector sizes of
512 (fdisk didn't tell me otherwise when I loaded the stick with it).  I
have build 542, q2c22 currently installed and have tried installing q2c25,
both on it's own (with 542 still being in the boot folder) and with the
newer build 558.  But have failed as described above.

Thanks for your help,
Philip

P.S Sorry for the grammar, it's really late
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.laptop.org/pipermail/devel/attachments/20070831/631b7897/attachment.htm 

Reply | Threaded
Open this post in threaded view
|

Unable to update my machine

Kimberley Quirk
Philip,

It sounds like you have tried a lot of things, but you don't mention
downloading the latest copy of olpc-auto.zip, which is required these days.
This file has been updated quite a few times recently so the latest version
should be downloaded whenever you get a new build.

The Autoreinstallation page of the wiki is very detailed and also has been
updated quite a bit in the last few weeks to help people with installation.
If you haven't read through that page carefully in the last few days, please
start there and follow the instructions carefully. Over time this will get
easier and it won't change with every build... but since we added the
ability to backup and restore and the activation key, it has taken a number
of weeks to get this process stabilized.

If you HAVE carefully read through the autoreinstallation page and you are
still having a problem, then we may need to see your machine or maybe we can
go through the installation together on irc.

Keep me posted on your progress,
Kim Quirk



On 8/30/07, Philip Macpherson <philipmac1 at gmail.com> wrote:

>
> To Devel list
>
> I've got a b4 which I've upgraded once when I first got it and am unable
> to update since.  I'm using the same usb flash key as I used the first
> time.  I first tried it unsuccessfully with all the files the same except
> for updated firmware .  Since then I've tried multiple times with the same
> files and later with the newer firmware and newer build.  I've also tried re
> downloading the files along with reformatting the stick (as both FAT 16 and
> 32)to see if it helps.
>
> When I insert the stick and turn on the machine with the x key pressed it
> recognises the key is there but doesn't seem to see the script.  As it
> doesn't  mention that it's either upgrading the machine or if it's firmware
> is current.
> The only difference I see after this happening is the usb stick now has a
> folder on it called .olpc.store
>
> The usb stick has only one partition and as far as I know sector sizes of
> 512 (fdisk didn't tell me otherwise when I loaded the stick with it).  I
> have build 542, q2c22 currently installed and have tried installing q2c25,
> both on it's own (with 542 still being in the boot folder) and with the
> newer build 558.  But have failed as described above.
>
> Thanks for your help,
> Philip
>
> P.S Sorry for the grammar, it's really late
>
> _______________________________________________
> Devel mailing list
> Devel at lists.laptop.org
> http://lists.laptop.org/listinfo/devel
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.laptop.org/pipermail/devel/attachments/20070830/12e36a9d/attachment-0001.htm 

Reply | Threaded
Open this post in threaded view
|

Unable to update my machine

crazy_bus
To Kim Quirk

Thanks for the reply.  My copy of olpc-auto.zip is 3 days old.  I last went
though the wiki page last night in a final attempt before emailing the
list.  I would really appreciate if you went though the installation with me
on irc as I've read though the page multiple times and can't see anything
I'm doing incorrectly.

Thanks again,
Philip

Here's a list of the files on the stick if it helps;
drwx------ 2 philip philip 16384 2007-08-28 17:26 boot
-rwx------ 1 philip philip 23623 2007-08-18 09:10 README

./boot:
total 242864
-rwx------ 1 philip philip     16731 2007-08-29 10:59 os558.crc
-rwx------ 1 philip philip 243662848 2007-08-29 10:58 os558.img
-rwx------ 1 philip philip     16876 2007-08-26 09:05 olpc.fth
-rwx------ 1 philip philip   1048576 2007-08-19 11:59 q2c25.rom
-rwx------ 1 philip philip   2333978 2007-08-18 02:04 usbupgrd.img
-rwx------ 1 philip philip   1559984 2007-08-16 06:00 usbupgos.img
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.laptop.org/pipermail/devel/attachments/20070831/7a3020b9/attachment.htm 

Reply | Threaded
Open this post in threaded view
|

Unable to update my machine

Eben Eliason
You may wish to try the python script that I recently added to the
autoreinstallation page.  It's been working consistently for me here.
The following should give you a completely clean auto-installation
image of the latest development build:

python sugar-update.py -vo /path/to/usbdrive

- Eben


On 8/30/07, Philip Macpherson <philipmac1 at gmail.com> wrote:

> To Kim Quirk
>
> Thanks for the reply.  My copy of olpc-auto.zip is 3 days old.  I last went
> though the wiki page last night in a final attempt before emailing the list.
>  I would really appreciate if you went though the installation with me on
> irc as I've read though the page multiple times and can't see anything I'm
> doing incorrectly.
>
> Thanks again,
> Philip
>
> Here's a list of the files on the stick if it helps;
> drwx------ 2 philip philip 16384 2007-08-28 17:26 boot
> -rwx------ 1 philip philip 23623 2007-08-18 09:10 README
>
> ./boot:
> total 242864
> -rwx------ 1 philip philip     16731 2007-08-29 10:59 os558.crc
> -rwx------ 1 philip philip 243662848 2007-08-29 10:58 os558.img
> -rwx------ 1 philip philip     16876 2007-08-26 09:05 olpc.fth
> -rwx------ 1 philip philip   1048576 2007-08-19 11:59 q2c25.rom
> -rwx------ 1 philip philip   2333978 2007-08-18 02:04 usbupgrd.img
> -rwx------ 1 philip philip   1559984 2007-08-16 06:00 usbupgos.img
>
>
> _______________________________________________
> Devel mailing list
> Devel at lists.laptop.org
> http://lists.laptop.org/listinfo/devel
>
>

Reply | Threaded
Open this post in threaded view
|

Unable to update my machine

crazy_bus
It now works.  Jerub on the wiki helped me with this.

[14:14] <Jerub> crazy_bus: I found just now that a drive with a partition
> table and a fat partition has that problem.
>
[14:14] <Jerub> but one with no partition taboe, and /dev/sda itself
> formatted as FAT, did not have that problem.


I formatted so I had no partition and then open firmware was able to read
it.  Maybe this should be mentioned on the wiki.

Thanks for all your help,
Philip
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.laptop.org/pipermail/devel/attachments/20070831/ece550b6/attachment.htm 

Reply | Threaded
Open this post in threaded view
|

Unable to update my machine

Mitch Bradley-2
The problem is ostensibly fixed in a new version of firmware that will
be released in the next couple of days.

The problem is actually a little more complicated than having a
partition table or not.  There was a situation where a disk could appear
to both have a partition table, and also have a FAT file system starting
in the first sector.  The way that can happen is if you use mkdosfs on
the raw disk (sda), and then change your mind and use fdisk to partition
it and then do mkdosfs again on sda1.  fdisk does not automatically
erase the BIOS parameter block (the FAT equivalent of a super block) in
the partition sector, so you end up with an ambiguous situation where
the first sector looks like an unpartitioned FAT volume and also a
partitioned volume at the same time.

I tried to correct that problem, but in the process I managed to break
the case where there is an unpartitioned FAT volume that has the
extended form of BIOS parameter block.

I think I have now fixed that breakage...  At least I hope so.

fdisk + mkdosfs is a mistake-prone combination.

Philip Macpherson wrote:

> It now works.  Jerub on the wiki helped me with this.
>
>     [14:14] <Jerub> crazy_bus: I found just now that a drive with a
>     partition table and a fat partition has that problem.
>
>     [14:14] <Jerub> but one with no partition taboe, and /dev/sda
>     itself formatted as FAT, did not have that problem.
>
>
> I formatted so I had no partition and then open firmware was able to
> read it.  Maybe this should be mentioned on the wiki.
>
> Thanks for all your help,
> Philip
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Devel mailing list
> Devel at lists.laptop.org
> http://lists.laptop.org/listinfo/devel
>  


Reply | Threaded
Open this post in threaded view
|

Unable to update my machine

Yuan Chao
On 8/31/07, Mitch Bradley <wmb at laptop.org> wrote:

Sorry to Mitch for multiple, I forgot to reply to "all".

> The problem is ostensibly fixed in a new version of firmware that will
> be released in the next couple of days.
That would be great.

I have a 2GB USB disk that used to be as a rescue boot disk. As
syslinux doesn't support >1GB, I need to partition it into two FAT16.
The olpc firmware doesn't seem to like it and every time I need to
specify the boot-device under OK prompt to do the update. However,
later I noticed that actually I only need to go to the OK prompt and
type boot would be sufficient. (USB icon only appears if boot from OK
prompt.) This makes me wonder that if the problem is due to the
partitioning (I'd prefer to keep my present partitions) or the USB
detection?


> The problem is actually a little more complicated than having a
> partition table or not.  There was a situation where a disk could appear
> to both have a partition table, and also have a FAT file system starting
> in the first sector.  The way that can happen is if you use mkdosfs on
> the raw disk (sda), and then change your mind and use fdisk to partition
> it and then do mkdosfs again on sda1.  fdisk does not automatically
> erase the BIOS parameter block (the FAT equivalent of a super block) in
> the partition sector, so you end up with an ambiguous situation where
> the first sector looks like an unpartitioned FAT volume and also a
> partitioned volume at the same time.
>
> I tried to correct that problem, but in the process I managed to break
> the case where there is an unpartitioned FAT volume that has the
> extended form of BIOS parameter block.
>
> I think I have now fixed that breakage...  At least I hope so.
>
> fdisk + mkdosfs is a mistake-prone combination.
>
> Philip Macpherson wrote:
> > It now works.  Jerub on the wiki helped me with this.
> >
> >     [14:14] <Jerub> crazy_bus: I found just now that a drive with a
> >     partition table and a fat partition has that problem.
> >
> >     [14:14] <Jerub> but one with no partition taboe, and /dev/sda
> >     itself formatted as FAT, did not have that problem.
> >
> >
> > I formatted so I had no partition and then open firmware was able to
> > read it.  Maybe this should be mentioned on the wiki.
> >
> > Thanks for all your help,
> > Philip
> >
> > ------------------------------------------------------------------------
> >
> > _______________________________________________
> > Devel mailing list
> > Devel at lists.laptop.org
> > http://lists.laptop.org/listinfo/devel
> >
>
> _______________________________________________
> Devel mailing list
> Devel at lists.laptop.org
> http://lists.laptop.org/listinfo/devel
>


--
Best regards,
Yuan Chao

Reply | Threaded
Open this post in threaded view
|

Unable to update my machine

Yuan Chao
In reply to this post by Mitch Bradley-2
On 8/31/07, Mitch Bradley <wmb at laptop.org> wrote:

> The fact that it does work from the ok prompt suggests that the problem
> is indeed the USB detection instead of the partitioning.  Is this a
Mmm... originally, according to the wiki and previous discussions,
most problem seems lead to partitioning. Maybe this slow detection
issue should worth listed on wiki?

> FLASH device or a real disk?  We have seen some USB disks that take
> several seconds to detect.
It's a cheap and slow USB 2.0 flash drive.

> When you hold down the game key to get the ok prompt, that gives the USB
> device a longer time between when USB power is applied and when
> enumeration starts - power is applied before the "Release the game key"
> message and enumeration happens afterwards, so you can control the time
> delay by holding the key longer.
> It should be possible to auto-boot by holding the key, then releasing
> it, then letting the countdown expire without typing the game key.
>
> You can also hold down a different game key, not the "X" one.  Any game
> key will trigger the "Release the game key" delay, but only the X key
> enables the OK prompt.
Ok. I'll try how it works.

> I am hesitant to increase the USB enumeration time delay, because that
> would penalize the majority of systems that will boot nearly always from
> NAND FLASH.
Sure.


--
Best regards,
Yuan Chao

Reply | Threaded
Open this post in threaded view
|

Unable to update my machine

Yuan Chao
In reply to this post by Mitch Bradley-2
On 8/31/07, Mitch Bradley <wmb at laptop.org> wrote:

> FLASH device or a real disk?  We have seen some USB disks that take
> several seconds to detect.
Ok. Now I can confirm that it's a slow detection problem in my case.
Either holding the game key for 2+ sec or let the count down expires
then boot will make it work.

> When you hold down the game key to get the ok prompt, that gives the USB
> device a longer time between when USB power is applied and when
> enumeration starts - power is applied before the "Release the game key"
> message and enumeration happens afterwards, so you can control the time
> delay by holding the key longer.
>
> It should be possible to auto-boot by holding the key, then releasing
> it, then letting the countdown expire without typing the game key.
>
> You can also hold down a different game key, not the "X" one.  Any game
> key will trigger the "Release the game key" delay, but only the X key
> enables the OK prompt.
Thanks for the explanation.


--
Best regards,
Yuan Chao

Reply | Threaded
Open this post in threaded view
|

Unable to update my machine

Hal Murray-2
In reply to this post by Mitch Bradley-2

Mitch said:

> fdisk + mkdosfs is a mistake-prone combination.

That's probably what I screwed up ages ago.

It might help if the wiki section had an example of something known to work -
that is take a USB flash drive in unknown state, erase everything, and set it
to a good state.

The corresponding mount command would probably be handy too.

I'm probably assuming that everybody is using a RedHat style linux, or
something close enough.  It may be worth recording the distribution (and
version) just in case that matters.

--
These are my opinions, not necessarily my employer's.  I hate spam.




Reply | Threaded
Open this post in threaded view
|

Unable to update my machine

James Cameron-2
On 02/09/2007, at 3:33 PM, Hal Murray wrote:
> It might help if the wiki section had an example of something known  
> to work -
> that is take a USB flash drive in unknown state, erase everything,  
> and set it
> to a good state.

Good idea.  Perhaps something that wiped the partition table and  
other stuff at the start of the drive.  If I knew exactly how the  
problem was caused, I'd recommend something, but I've missed the  
reproducer.  I admit, a full erase did fix a problem like this for me  
earlier last month.

--
James Cameron