UNSOLVED - Remaster on usb stick gives "boot error"

[Maurice]

Maurice

On this USB install, we just need to make sure it has the boot flag.  This does not destroy anything.

root konsole fdisk /dev/sdb

don't use the m, well you can no harm
a
1
w
q

fdisk-l and post so I can see the boot flag has gone up.  Same as you have done before.

Just try and boot the DG33FB.

The next step is to kick the geometry, instructions next post

fdisk /dev/sdb

Command (m for help): a
Partition number (1-4): 1

Command (m for help): w
The partition table has been altered!

Calling ioctl() to re-read partition table.

WARNING: Re-reading the partition table failed with error 16: Device or resource busy.
The kernel still uses the old table. The new table will be used at
the next reboot or after you run partprobe( or kpartx(
Syncing disks.


fdisk -l

Disk /dev/sdb: 16.0 GB, 16039018496 bytes
255 heads, 63 sectors/track, 1949 cylinders, total 31326208 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x0009869e

   Device Boot      Start         End      Blocks   Id  System
/dev/sdb1   *          63     2216969     1108453+   6  FAT16
/dev/sdb2         2216970    31310684    14546857+  83  Linux


Edit

DG33FB - Not a bootable disk

[Old-Polack]

Old-Polack

I wondered how long I would get away with it, let us say no I did not know that.  

May I ask you, can you see any reason why the bootflag can have any impact when we know there is a grub on the USB key. I am aware of this seems to have solved problems but struggle to see why.

I fail to see why a boot flag would have any affect, with grub installed to the MBR, unless the BIOS itself is buggy. The only time I've toggled the boot flag on has been to demonstrate how it's done for someone dual booting with some version of Windows, and to show that grub, installed to the boot sector of a primary partition holding a Linux installation, will, in fact, boot with the MS boot code in the MBR, if the boot flag is toggled on for that primary partition .

Considering all the trouble this USB stick is having with this particular machine, if Maurice has a way to install the MS boot code to the MBR of the stick, it might be worth a try, doing that, installing grub to the / partition boot sector, and toggling the boot flag on, for that partition. 

[wedgetail]

Maurice

Well done, so we have proved that it was not the boot flag.  I have a number of avenues to follow but with reference to what Old-Polack tells us I have for some time suspected the BIOS so we will just make a detour here and see if you have the latest installed;  Simple little exercise, in konsole do this

dmesg | grep BIOS

Will look something like that

DMI: System manufacturer System Product Name/P5P41D, BIOS 0207    07/21/2009
PCI: PCI BIOS revision 3.00 entry at 0xf0031, last bus=5
PnPBIOS: Disabled
apm: BIOS version 1.2 Flags 0x03 (Driver version 1.16ac)
BIOS EDD facility v0.16 2004-Jun-25, 3 devices found
hda_codec: ALC887: BIOS auto-probing.
[root@localhost gert]#

Old-Polack
Is it enough just to dd in the bytes up to bottom of the partition table?

(will add a bit here soon)

[Maurice]

Maurice

Well done, so we have proved that it was not the boot flag.  I have a number of avenues to follow but with reference to what Old-Polack tells us I have for some time suspected the BIOS so we will just make a detour here and see if you have the latest installed;  Simple little exercise, in konsole do this

dmesg | grep BIOS

dmesg | grep BIOS
BIOS-provided physical RAM map:
 BIOS-e820: 0000000000000000 - 000000000009fc00 (usable)
 BIOS-e820: 000000000009fc00 - 00000000000a0000 (reserved)
 BIOS-e820: 00000000000e0000 - 0000000000100000 (reserved)
 BIOS-e820: 0000000000100000 - 00000000ce76f000 (usable)
 BIOS-e820: 00000000ce76f000 - 00000000ce830000 (ACPI NVS)
 BIOS-e820: 00000000ce830000 - 00000000cfac6000 (usable)
 BIOS-e820: 00000000cfac6000 - 00000000cfac8000 (reserved)
 BIOS-e820: 00000000cfac8000 - 00000000cfb8d000 (usable)
 BIOS-e820: 00000000cfb8d000 - 00000000cfbe5000 (ACPI NVS)
 BIOS-e820: 00000000cfbe5000 - 00000000cfbea000 (usable)
 BIOS-e820: 00000000cfbea000 - 00000000cfbf3000 (ACPI data)
 BIOS-e820: 00000000cfbf3000 - 00000000cfbf4000 (usable)
 BIOS-e820: 00000000cfbf4000 - 00000000cfbff000 (ACPI data)
 BIOS-e820: 00000000cfbff000 - 00000000cfc00000 (usable)
 BIOS-e820: 00000000cfc00000 - 00000000d0000000 (reserved)
 BIOS-e820: 00000000f0000000 - 00000000f8000000 (reserved)
 BIOS-e820: 00000000fff00000 - 0000000100000000 (reserved)
 BIOS-e820: 0000000100000000 - 000000012c000000 (usable)
  #0 [0000000000 - 0000001000]   BIOS data page ==> [0000000000 - 0000001000]
  #4 [000009fc00 - 00000fe200]    BIOS reserved ==> [000009fc00 - 00000fe200]
  #6 [00000fe250 - 0000100000]    BIOS reserved ==> [00000fe250 - 0000100000]
PnPBIOS: Disabled
apm: BIOS version 1.2 Flags 0x03 (Driver version 1.16ac)
BIOS EDD facility v0.16 2004-Jun-25, 1 devices found

[wedgetail]

Maurice
Ooops, did not work not sure where we can find your BIOS info. 

[djohnston]

wedgetail,

Not sure what you're looking for or at. There's a lot of BIOS messages listed right there. If you want the full specs of BIOS, biosdecode and dmidecode run as root are the tools. But, I think you're missing two key components.

PnPBIOS: Disabled

BIOS EDD facility v0.16 2004-Jun-25, 1 devices found


Plug'nPlay is turned off. Really should be on.
He has the newer EDD. Look here for a description.

[Old-Polack]

Old-Polack
Is it enough just to dd in the bytes up to bottom of the partition table?

(will add a bit here soon)

Need to have the bytes saved somewhere first, like a saved file. Assuming one has a booting Windows drive, and it's /dev/sdb, as seen from a liveCD

[root@localhost ~]# dd if=/dev/sdb of=/path/to/file/ms-boot-code.bak bs=446 count=1

then still as root, assuming the USB stick shows as /dev/sdc;

[root@localhost ~]# dd if=/path/to/file/ms-boot-code.bak of=/dev/sdc bs=446 count=1

One could copy from the hard drive MBR directly to the USB MBR, but this way one has a backup file, in case one needs to do this again, and there's no Windows drive around.

[Maurice]

Maurice
Ooops, did not work not sure where we can find your BIOS info. 

I'm a little ahead of you.    When you mentioned bios I remembered that my version was a bit ancient so started looking for an update. I found the correct version and downloaded the bios iso file from Intel and will install it shortly. - with digits tightly crossed.

[wedgetail]

Old-Polack
Fully understood.  Perhaps better come back to this one later, dd is powerful 

/dev/sdb1   *          63     2216969     1108453+   6  FAT16
/dev/sdb2         2216970    31310684    14546857+  83  Linux

The + at the end of the lines, can't find meaning but I thought it has something to do with not quite on the cylinders however above two lines suffer slightly if each track has 63 sectors.

If I insist on proper "63" alignment beginning and end what figures would you allocate.  The start of both partitions are multiple of "63" but the endpoint is not and therfore the number of blocks are not multiples of "63" either in this case.

The reason I mention this is that the 3 USB drives Just18 has working I think it was two of them were resurrected. After resurrection these two have almost perfect track alignment. 

On Maurice's existing unit it would be possible just to shorten the partitions a bit and try?  What is your opinion

djohnston

Thanks, I am not sure of the two items you picked up but I would prefer to leave them at present.  What I am aiming at is something that looks different, I will have to be more specific but it got lost in my notes.  The thing is that upgrading the BIOS did sort a USB problem out in one case on this particular motherboard.  I have been half assuming that this DG33BF way younger and would have latest BIOS.  Also it is a bit of an operation to change that, all tools are available from Intel.  I was leaving this one until other options tried.   

I have analysed Just18 data and I thought I found a pattern there and that he had the solution. So I have tried to follow his path to the bitter end. I have carried out experiments on one of my problem drives ( I have two but it is another problem) last night using diskdrake, gparted , partionmanager and fdisk.  I am still not sure what to think of what I have seen.

Maurice

I am crossing both legs and fingers and knocking under the table (it is wood)

[Maurice]

DG33FB----Not a bootable disk...... So relax DJ and stand up straight again.

I updated bios with Success and expected to see a shiny new more modern looking bios than the old fashioned looking thing I had before but, seemingly, nothing had changed. Shouldn't there have been some visible change?

However - the same negative result made me wonder if it is worth carrying on this thread. In spite of your valiant efforts - especially wedgetail's - it seems to be one of those intractable problems without solution. If you all want to call it off I don't blame you and I'll understand.

[Old-Polack]

wedgetail:

Don't worry about the +. Try to get rid of them and you'll do more damage than good.

[root@fatman ~]# fdisk -l

Code: [Select]

Disk /dev/sdb: 750.2 GB, 750156374016 bytes
255 heads, 63 sectors/track, 91201 cylinders, total 1465149168 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x2db1883a

   Device Boot      Start         End      Blocks   Id  System
/dev/sdb1              63      208844      104391   83  Linux
/dev/sdb2          208845    12225464     6008310   82  Linux swap / Solaris
/dev/sdb3        12225465    72244304    30009420   83  Linux
/dev/sdb4        72244305  1465144064   696449880    5  Extended
/dev/sdb5        72244368   272253554   100004593+  83  Linux
/dev/sdb6       272253618   670713749   199230066   83  Linux
/dev/sdb7       670713813   731278799    30282493+  83  Linux
/dev/sdb8       731278863   811378889    40050013+  83  Linux
/dev/sdb9       811378953  1045767239   117194143+  83  Linux
/dev/sdb10     1045767303  1280155589   117194143+  83  Linux
/dev/sdb11     1280155653  1465144064    92494206   83  Linux

Disk /dev/sda: 1000.2 GB, 1000204886016 bytes
255 heads, 63 sectors/track, 121601 cylinders, total 1953525168 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x0000d21c

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1              63      626534      313236   83  Linux
/dev/sda2          626535    16820054     8096760   82  Linux swap / Solaris
/dev/sda3        16820055   114495254    48837600   83  Linux
/dev/sda4       114495255  1953520064   919512405    5  Extended
/dev/sda5       114495318   329332499   107418591   83  Linux
/dev/sda6       329332563   534434354   102550896   83  Linux
/dev/sda7       534434418   596943269    31254426   83  Linux
/dev/sda8       596943333   659452184    31254426   83  Linux
/dev/sda9       659452248   721961099    31254426   83  Linux
/dev/sda10      721961163   784470014    31254426   83  Linux
/dev/sda11      784470078   994198589   104864256   83  Linux
/dev/sda12      994198653  1623368249   314584798+  83  Linux
/dev/sda13     1623368313  1685893229    31262458+  83  Linux
/dev/sda14     1685893293  1749366044    31736376   83  Linux
/dev/sda15     1749366108  1816485614    33559753+  83  Linux
/dev/sda16     1816487663  1889887982    36700160   83  Linux
/dev/sda17     1889892080  1953520064    31813992+  83  Linux

Doesn't hurt a thing.

[wedgetail]

Old-Polack

On a "real" drive I have just looked at them as a curio, but now when we are talking about tables/figures only or not real cylinders I am a bit more apprehensive. I understand what you mean here.

Are you implying that the length of a partition in general can end just about anywhere as long as the start is 'lined' up?

Be very interesting to see what Maurice comes up with when new BIOS installed.  

[Old-Polack]

Old-Polack

On a "real" drive I have just looked at them as a curio, but now when we are talking about tables/figures only or not real cylinders I am a bit more apprehensive. I understand what you mean here.

Are you implying that the length of a partition in general can end just about anywhere as long as the start is 'lined' up?

Be very interesting to see what Maurice comes up with when new BIOS installed.  

Yes!

[wedgetail]

Maurice

No, I am not quite done yet.  BIOS changes I don't think normally shows any external signs an upgrade in that respect is not very interesting.

Do you happen to have a real external USB hard drive?  Not a solid state drive.

[Maurice]

Maurice

No, I am not quite done yet.  BIOS changes I don't think normally shows any external signs an upgrade in that respect is not very interesting.

Do you happen to have a real external USB hard drive?  Not a solid state drive.

No

I looked at the bios again and there was a change - in the 2008 to 2009 so it was not as old as I thought it was.