• Data recovery via USB?

    From Philip Herlihy@21:1/5 to All on Thu Sep 30 12:07:09 2021
    A friend has asked me to recover photos from the system disk of an ancient PC. I booted the old PC, and the disk is in need of "initialisation". Now, my knowledge at this level is pretty hazy, but I figure I'll need to use something Testdisk to recover the partition, and hopefully its contents.

    But it's an IDE disk. I do have an IDE to USB3 adapter, but I've always suspected that the USB interface "gets in the way" of the low-level activity likely to be involved in this sort of recovery. I've found in the past that USB2 adapters don't allow SMART data to be read, for example, and only sometimes (?) has this been possible with USB3 adapters.

    I do have an old PC (grindingly slow) which has IDE ports on the board, but it would be much more convenient to do this work on a newer, faster machine. My principal PC has an E-SATA card which I've found very useful for attaching a SATA disk dock - it's just like the disks are directly connected to the board.

    So would it make sense to buy an E-SATA to SATA adapter cable, plus a SATA to IDE adapter? Or would the 'adapters' somehow mangle things (as the USB2 adapters seemed to do)? I'll be grateful for any advice from the smart folk here.


    --

    Phil, London

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From newshound@21:1/5 to Philip Herlihy on Thu Sep 30 13:21:30 2021
    On 30/09/2021 12:07, Philip Herlihy wrote:
    A friend has asked me to recover photos from the system disk of an ancient PC.
    I booted the old PC, and the disk is in need of "initialisation". Now, my knowledge at this level is pretty hazy, but I figure I'll need to use something
    Testdisk to recover the partition, and hopefully its contents.

    But it's an IDE disk. I do have an IDE to USB3 adapter, but I've always suspected that the USB interface "gets in the way" of the low-level activity likely to be involved in this sort of recovery. I've found in the past that USB2 adapters don't allow SMART data to be read, for example, and only sometimes (?) has this been possible with USB3 adapters.

    I do have an old PC (grindingly slow) which has IDE ports on the board, but it
    would be much more convenient to do this work on a newer, faster machine. My principal PC has an E-SATA card which I've found very useful for attaching a SATA disk dock - it's just like the disks are directly connected to the board.

    So would it make sense to buy an E-SATA to SATA adapter cable, plus a SATA to IDE adapter? Or would the 'adapters' somehow mangle things (as the USB2 adapters seemed to do)? I'll be grateful for any advice from the smart folk here.


    Can't really answer the question, but personally I would start by seeing
    what your IDE machine can recover. Do you think that is going to require continuous intervention on your part? To my mind there is more to go
    wrong as you introduce interfaces.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From jkn@21:1/5 to Philip Herlihy on Thu Sep 30 08:02:25 2021
    On Thursday, September 30, 2021 at 12:07:10 PM UTC+1, Philip Herlihy wrote:
    A friend has asked me to recover photos from the system disk of an ancient PC.
    I booted the old PC, and the disk is in need of "initialisation". Now, my knowledge at this level is pretty hazy, but I figure I'll need to use something
    Testdisk to recover the partition, and hopefully its contents.

    But it's an IDE disk. I do have an IDE to USB3 adapter, but I've always suspected that the USB interface "gets in the way" of the low-level activity likely to be involved in this sort of recovery. I've found in the past that USB2 adapters don't allow SMART data to be read, for example, and only sometimes (?) has this been possible with USB3 adapters.

    I do have an old PC (grindingly slow) which has IDE ports on the board, but it
    would be much more convenient to do this work on a newer, faster machine. My principal PC has an E-SATA card which I've found very useful for attaching a SATA disk dock - it's just like the disks are directly connected to the board.

    So would it make sense to buy an E-SATA to SATA adapter cable, plus a SATA to IDE adapter? Or would the 'adapters' somehow mangle things (as the USB2 adapters seemed to do)? I'll be grateful for any advice from the smart folk here.

    Depends a bit on your level of knowledge, but could you:

    - boot the ancient PC from a USB drive running Linux, and maybe with another partition on it
    - image the suspect drive onto the USB drive
    - take the contents away and look at them at leisure with TestDisk
    ?

    J^n

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Philip Herlihy@21:1/5 to All on Thu Sep 30 15:17:54 2021
    In article <kb-dnbYbyKhXNsj8nZ2dnUU78SHNnZ2d@brightview.co.uk>, newshound@stevejqr.plus.com says...

    On 30/09/2021 12:07, Philip Herlihy wrote:
    A friend has asked me to recover photos from the system disk of an ancient PC.
    I booted the old PC, and the disk is in need of "initialisation". Now, my knowledge at this level is pretty hazy, but I figure I'll need to use something
    Testdisk to recover the partition, and hopefully its contents.

    But it's an IDE disk. I do have an IDE to USB3 adapter, but I've always suspected that the USB interface "gets in the way" of the low-level activity
    likely to be involved in this sort of recovery. I've found in the past that
    USB2 adapters don't allow SMART data to be read, for example, and only sometimes (?) has this been possible with USB3 adapters.

    I do have an old PC (grindingly slow) which has IDE ports on the board, but it
    would be much more convenient to do this work on a newer, faster machine. My
    principal PC has an E-SATA card which I've found very useful for attaching a
    SATA disk dock - it's just like the disks are directly connected to the board.

    So would it make sense to buy an E-SATA to SATA adapter cable, plus a SATA to
    IDE adapter? Or would the 'adapters' somehow mangle things (as the USB2 adapters seemed to do)? I'll be grateful for any advice from the smart folk
    here.


    Can't really answer the question, but personally I would start by seeing
    what your IDE machine can recover. Do you think that is going to require continuous intervention on your part? To my mind there is more to go
    wrong as you introduce interfaces.

    There's a practical disincentive to use that machine. I'd need to haul it out of its present niche, and work with it opened up in a household with little free space. It's also wearyingly slow. If the interfaces I've touched upon work as well as the E-SATA docking station I use for SATA disks then there's real benefit in using them.

    --

    Phil, London

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From David@21:1/5 to Philip Herlihy on Thu Sep 30 15:32:49 2021
    On Thu, 30 Sep 2021 12:07:09 +0100, Philip Herlihy wrote:

    A friend has asked me to recover photos from the system disk of an
    ancient PC. I booted the old PC, and the disk is in need of
    "initialisation". Now, my knowledge at this level is pretty hazy, but I figure I'll need to use something Testdisk to recover the partition, and hopefully its contents.

    But it's an IDE disk. I do have an IDE to USB3 adapter, but I've always suspected that the USB interface "gets in the way" of the low-level
    activity likely to be involved in this sort of recovery. I've found in
    the past that USB2 adapters don't allow SMART data to be read, for
    example, and only sometimes (?) has this been possible with USB3
    adapters.

    I do have an old PC (grindingly slow) which has IDE ports on the board,
    but it would be much more convenient to do this work on a newer, faster machine. My principal PC has an E-SATA card which I've found very
    useful for attaching a SATA disk dock - it's just like the disks are
    directly connected to the board.

    So would it make sense to buy an E-SATA to SATA adapter cable, plus a
    SATA to IDE adapter? Or would the 'adapters' somehow mangle things (as
    the USB2 adapters seemed to do)? I'll be grateful for any advice from
    the smart folk here.

    <https://www.amazon.co.uk/Dynamode-Port-Sata-Card-Electronics/dp/
    B001A50LH6/>

    Would something like this do the trick?
    About the same price as the external USB adapters and the eSATA adapters,
    I think.

    I have no idea if this is a true IDE card (with drivers) or some bodge to convert to SATA on the card.

    Cheers


    Dave R

    --
    AMD FX-6300 in GA-990X-Gaming SLI-CF running Windows 7 Pro x64

    --
    This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Theo@21:1/5 to Philip Herlihy on Thu Sep 30 17:07:50 2021
    Philip Herlihy <thiswillbounceback@you.com> wrote:
    A friend has asked me to recover photos from the system disk of an ancient PC. I booted the old PC, and the disk is in need of "initialisation".
    Now, my knowledge at this level is pretty hazy, but I figure I'll need to
    use something Testdisk to recover the partition, and hopefully its
    contents.

    But it's an IDE disk. I do have an IDE to USB3 adapter, but I've always suspected that the USB interface "gets in the way" of the low-level
    activity likely to be involved in this sort of recovery. I've found in
    the past that USB2 adapters don't allow SMART data to be read, for
    example, and only sometimes (?) has this been possible with USB3 adapters.

    How ancient is it? If it's ~1995 or later it's likely to be using LBA addressing, which is what everything today uses. If it uses the older CHS addressing the adapter may talk to it incorrectly. I'm not sure when that
    died out in Windowsland, but maybe early 90s.

    I would just give it a try and see how you get on.

    If it's really old (PC-XT or something) the drive may only have an 8 bit connection, but that's pretty unlikely.

    SMART is a different command set that the adapter needs to explicitly
    support, and which needs to be translated between ATA and USB - many
    adapters don't handle that. It's not part of the mainline data
    communication interface.

    I do have an old PC (grindingly slow) which has IDE ports on the board,
    but it would be much more convenient to do this work on a newer, faster machine. My principal PC has an E-SATA card which I've found very useful
    for attaching a SATA disk dock - it's just like the disks are directly connected to the board.

    So would it make sense to buy an E-SATA to SATA adapter cable, plus a SATA
    to IDE adapter? Or would the 'adapters' somehow mangle things (as the
    USB2 adapters seemed to do)? I'll be grateful for any advice from the
    smart folk here.

    I'd have thought a SATA to PATA adapter is at the same risk as mangling
    things as a PATA to USB adapter.

    Does your old PC have a USB card? Or can it take a PCI one? If the
    USB-PATA adapter doesn't work, I'd try using it as an intermediary to copy
    from PATA to a USB stick.

    If you can boot a Linux 'recovery CD' on it, taking a raw drive image
    would be something like:

    dd if=/dev/sda of=rawdisk.img bs=512

    assuming you made sure that sda was referring to the right drive.
    Or:

    dd if=/dev/sda of=/dev/sdb bs=512

    if you wanted to clone to another (USB?) drive.

    Theo

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Theo@21:1/5 to David on Thu Sep 30 17:12:06 2021
    David <wibble@btinternet.com> wrote:
    <https://www.amazon.co.uk/Dynamode-Port-Sata-Card-Electronics/dp/ B001A50LH6/>

    Would something like this do the trick?
    About the same price as the external USB adapters and the eSATA adapters,
    I think.

    I have no idea if this is a true IDE card (with drivers) or some bodge to convert to SATA on the card.

    VT6421A chip, which has two SATA and one IDE port.

    So that would do, assuming you have a PCI (not PCIe) slot. As long as the drive isn't so old (80s) that current software won't talk to it.

    Theo

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Philip Herlihy@21:1/5 to All on Fri Oct 1 09:27:53 2021
    In article <64575f17-c7c3-44d6-b9f8-c1d5422b5a1an@googlegroups.com>, jkn_gg@nicorp.f9.co.uk says...

    On Thursday, September 30, 2021 at 12:07:10 PM UTC+1, Philip Herlihy wrote:
    A friend has asked me to recover photos from the system disk of an ancient PC.
    I booted the old PC, and the disk is in need of "initialisation". Now, my knowledge at this level is pretty hazy, but I figure I'll need to use something
    Testdisk to recover the partition, and hopefully its contents.

    But it's an IDE disk. I do have an IDE to USB3 adapter, but I've always suspected that the USB interface "gets in the way" of the low-level activity
    likely to be involved in this sort of recovery. I've found in the past that USB2 adapters don't allow SMART data to be read, for example, and only sometimes (?) has this been possible with USB3 adapters.

    I do have an old PC (grindingly slow) which has IDE ports on the board, but it
    would be much more convenient to do this work on a newer, faster machine. My
    principal PC has an E-SATA card which I've found very useful for attaching a
    SATA disk dock - it's just like the disks are directly connected to the board.

    So would it make sense to buy an E-SATA to SATA adapter cable, plus a SATA to
    IDE adapter? Or would the 'adapters' somehow mangle things (as the USB2 adapters seemed to do)? I'll be grateful for any advice from the smart folk here.

    Depends a bit on your level of knowledge, but could you:

    - boot the ancient PC from a USB drive running Linux, and maybe with another partition on it
    - image the suspect drive onto the USB drive
    - take the contents away and look at them at leisure with TestDisk
    ?

    J^n

    I suspect the ancient PC is even more grindingly slow than my IDE-equipped one! I'm also generally distrustful of USB drives, and while I was once designated a Unix "guru", I haven't used anything Unix/Linux since Windows NT 4.1 was new! So interesting suggestion, but I think I'd go with the IDE-equipped PC (running W10) first. Thanks, of course.

    --

    Phil, London

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Philip Herlihy@21:1/5 to All on Fri Oct 1 09:34:54 2021
    In article <irm3l1F9jskU3@mid.individual.net>, wibble@btinternet.com says...

    On Thu, 30 Sep 2021 12:07:09 +0100, Philip Herlihy wrote:

    A friend has asked me to recover photos from the system disk of an
    ancient PC. I booted the old PC, and the disk is in need of "initialisation". Now, my knowledge at this level is pretty hazy, but I figure I'll need to use something Testdisk to recover the partition, and hopefully its contents.

    But it's an IDE disk. I do have an IDE to USB3 adapter, but I've always suspected that the USB interface "gets in the way" of the low-level activity likely to be involved in this sort of recovery. I've found in
    the past that USB2 adapters don't allow SMART data to be read, for
    example, and only sometimes (?) has this been possible with USB3
    adapters.

    I do have an old PC (grindingly slow) which has IDE ports on the board,
    but it would be much more convenient to do this work on a newer, faster machine. My principal PC has an E-SATA card which I've found very
    useful for attaching a SATA disk dock - it's just like the disks are directly connected to the board.

    So would it make sense to buy an E-SATA to SATA adapter cable, plus a
    SATA to IDE adapter? Or would the 'adapters' somehow mangle things (as
    the USB2 adapters seemed to do)? I'll be grateful for any advice from
    the smart folk here.

    <https://www.amazon.co.uk/Dynamode-Port-Sata-Card-Electronics/dp/ B001A50LH6/>

    Would something like this do the trick?
    About the same price as the external USB adapters and the eSATA adapters,
    I think.

    I have no idea if this is a true IDE card (with drivers) or some bodge to convert to SATA on the card.

    Cheers


    Dave R

    --
    AMD FX-6300 in GA-990X-Gaming SLI-CF running Windows 7 Pro x64

    Thaks. It certainly crossed my mind to look for something like this. I'd have to figure out which card to displace, though that's not a big deal. But I'm hoping that some of the folk here with encyclopaedic knowledge of the "low level" stuff will be able to clarify if the adapters would mangle things.

    --

    Phil, London

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Philip Herlihy@21:1/5 to All on Fri Oct 1 10:38:07 2021
    In article <cPf*1Gxvy@news.chiark.greenend.org.uk>, theom+news@chiark.greenend.org.uk says...

    Philip Herlihy <thiswillbounceback@you.com> wrote:
    A friend has asked me to recover photos from the system disk of an ancient PC. I booted the old PC, and the disk is in need of "initialisation".
    Now, my knowledge at this level is pretty hazy, but I figure I'll need to use something Testdisk to recover the partition, and hopefully its contents.

    But it's an IDE disk. I do have an IDE to USB3 adapter, but I've always suspected that the USB interface "gets in the way" of the low-level activity likely to be involved in this sort of recovery. I've found in
    the past that USB2 adapters don't allow SMART data to be read, for
    example, and only sometimes (?) has this been possible with USB3 adapters.

    How ancient is it? If it's ~1995 or later it's likely to be using LBA addressing, which is what everything today uses. If it uses the older CHS addressing the adapter may talk to it incorrectly. I'm not sure when that died out in Windowsland, but maybe early 90s.

    I would just give it a try and see how you get on.

    If it's really old (PC-XT or something) the drive may only have an 8 bit connection, but that's pretty unlikely.

    SMART is a different command set that the adapter needs to explicitly support, and which needs to be translated between ATA and USB - many
    adapters don't handle that. It's not part of the mainline data
    communication interface.

    I do have an old PC (grindingly slow) which has IDE ports on the board,
    but it would be much more convenient to do this work on a newer, faster machine. My principal PC has an E-SATA card which I've found very useful for attaching a SATA disk dock - it's just like the disks are directly connected to the board.

    So would it make sense to buy an E-SATA to SATA adapter cable, plus a SATA to IDE adapter? Or would the 'adapters' somehow mangle things (as the
    USB2 adapters seemed to do)? I'll be grateful for any advice from the smart folk here.

    I'd have thought a SATA to PATA adapter is at the same risk as mangling things as a PATA to USB adapter.

    Does your old PC have a USB card? Or can it take a PCI one? If the
    USB-PATA adapter doesn't work, I'd try using it as an intermediary to copy from PATA to a USB stick.

    If you can boot a Linux 'recovery CD' on it, taking a raw drive image
    would be something like:

    dd if=/dev/sda of=rawdisk.img bs=512

    assuming you made sure that sda was referring to the right drive.
    Or:

    dd if=/dev/sda of=/dev/sdb bs=512

    if you wanted to clone to another (USB?) drive.

    Theo

    Thanks. I don't know how old the original PC was - all I have now is the disk itself. It's a seagate ST3491A, stamped with a date of January 1995.

    I was particularly interested in what you said about SMART, which may or not be 'translated' by adapters. I guess what I'm fumbling to understand is what translation goes on. Clearly, in something trivial like an E-SATA to SATA cable/adapter it's just a case of mapping wires. But with a USB interface there is 'electronics' in there. What's it doing? What can't it do?

    Of course the difficulty with a data recovery task is the uncertainty over whether the disk is the problem or the interface you're using to attempt to access it. With a healthy disk, if you can see the files and maybe create a new one then you know where you are. With a disk that is seen as uninitialised - not so much!

    Current thinking is to try a PATA to SATA adapter and proceed carefully. I've found that if Acronis True Image can 'see' a disk, then an image can subsequently be 'mounted' and the files read - that's all read-only as far as the source disk is concerned. I might see if Testdisk can see the disk via such an adapter, but it's possibly more of a gamble to ask Testdisk to 'correct' the partition. I may just end up with my old IDE-capable PC all over the kitchen (and the missus all over me...).

    I keep telling myself I must get on top of Linux. But to operate safely, especially using a 'foreign' filesystem you do have to have a certain minimum of contextual understanding or you're going to screw something up. That might be relatively quickly achieved, especially for someone with a (historic) background in Unix (SunOS, anyone?) but it'll take more time than I currently have. Competent folk often lose sight of the breadth of their knowledge!

    --

    Phil, London

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From David@21:1/5 to Philip Herlihy on Fri Oct 1 10:26:40 2021
    On Fri, 01 Oct 2021 10:38:07 +0100, Philip Herlihy wrote:

    <snip>

    I keep telling myself I must get on top of Linux. But to operate
    safely, especially using a 'foreign' filesystem you do have to have a
    certain minimum of contextual understanding or you're going to screw something up. That might be relatively quickly achieved, especially for someone with a (historic) background in Unix (SunOS, anyone?) but it'll
    take more time than I currently have. Competent folk often lose sight
    of the breadth of their knowledge!

    IIRC SunOS was Berkley Systems Division Unix.
    [As opposed to AT&T System V.)]

    In which case a FreeBSD variant might be a good place to start.

    Mind you I've forgotten about 90% of my Unix knowledge by now.

    Like riding a bicycle it will probably come back quite quickly, but not
    the best route for a critical action as a learning experience.

    Cheers


    Dave R


    --
    AMD FX-6300 in GA-990X-Gaming SLI-CF running Windows 7 Pro x64

    --
    This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Philip Herlihy@21:1/5 to All on Fri Oct 1 12:26:13 2021
    In article <iro630F9jskU8@mid.individual.net>, wibble@btinternet.com says...

    On Fri, 01 Oct 2021 10:38:07 +0100, Philip Herlihy wrote:

    <snip>

    I keep telling myself I must get on top of Linux. But to operate
    safely, especially using a 'foreign' filesystem you do have to have a certain minimum of contextual understanding or you're going to screw something up. That might be relatively quickly achieved, especially for someone with a (historic) background in Unix (SunOS, anyone?) but it'll take more time than I currently have. Competent folk often lose sight
    of the breadth of their knowledge!

    IIRC SunOS was Berkley Systems Division Unix.
    [As opposed to AT&T System V.)]

    In which case a FreeBSD variant might be a good place to start.

    Mind you I've forgotten about 90% of my Unix knowledge by now.

    Like riding a bicycle it will probably come back quite quickly, but not
    the best route for a critical action as a learning experience.


    Just as an expert understanding of Windows 3.1 isn't likely to be much use in 2021!

    --

    Phil, London

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Theo@21:1/5 to Philip Herlihy on Fri Oct 1 17:41:44 2021
    Philip Herlihy <thiswillbounceback@you.com> wrote:
    In article <cPf*1Gxvy@news.chiark.greenend.org.uk>, theom+news@chiark.greenend.org.uk says...

    Thanks. I don't know how old the original PC was - all I have now is the disk
    itself. It's a seagate ST3491A, stamped with a date of January 1995.

    Do you know what OS it was running? That drive is about the borderline
    between when software and drives switched from CHS (cylinder/head/sector) to LBA (logical block addressing). It is possible current software and
    adapters don't handle CHS correctly, although it may be that the disc is equally readable in either format.

    If the original driver used the CHS in the wrong order you can end up with a permutation of your disc blocks, which makes the drive appear to contain garbage. With LBA there's no ambiguity.

    I was particularly interested in what you said about SMART, which may or not be
    'translated' by adapters. I guess what I'm fumbling to understand is what translation goes on. Clearly, in something trivial like an E-SATA to SATA cable/adapter it's just a case of mapping wires. But with a USB interface there is 'electronics' in there. What's it doing? What can't it do?

    USB Mass Storage is based on the SCSI protocol. IDE and SATA are based on
    the ATA protocol. It's a bit like English and French, and the adapter has
    to understand both fully to translate between the two. If you don't have
    the vocab to understand what the Frenchman says, you can't say the English equivalent - which is what happens with SMART.

    (There are ways to say 'just quote this French phrase and he'll understand' which cover some of the more obscure corners, but it still needs somebody to speak French)

    The trouble with adapters is the firmware running on them is not always as comprehensive or fully-featured as you might hope, since the chips are being knocked out for $1 a time.

    My concern is the CHS v LBA thing might blindside the USB adapter - while a native Frenchman might understand Norman French, somebody who barely speaks modern French might not.

    Current thinking is to try a PATA to SATA adapter and proceed carefully. I've
    found that if Acronis True Image can 'see' a disk, then an image can subsequently be 'mounted' and the files read - that's all read-only as far as the source disk is concerned. I might see if Testdisk can see the disk via such an adapter, but it's possibly more of a gamble to ask Testdisk to 'correct' the partition. I may just end up with my old IDE-capable PC all over
    the kitchen (and the missus all over me...).

    I'm not sure to what extent the cylinder/head/sector registers are
    replicated in SATA, or whether the interface is entirely LBA by that point.
    But I don't think it would be too dangerous to try - worst case it won't be detected, or appear as garbage.

    I would not ever think of writing to the disk, you can't be sure that won't damage something and it's your only copy. Get a bit image of it that you
    can duplicate, and then work on a copy of that (if you like, write it to a
    USB stick so it's easy to mount etc. Or connect it to a VM as a virtual
    disc)

    I keep telling myself I must get on top of Linux. But to operate safely, especially using a 'foreign' filesystem you do have to have a certain minimum of contextual understanding or you're going to screw something up. That might
    be relatively quickly achieved, especially for someone with a (historic) background in Unix (SunOS, anyone?) but it'll take more time than I currently have. Competent folk often lose sight of the breadth of their knowledge!

    Agreed. I'm not sure what the best user-friendly danger-free recovery
    toolkit is, perhaps there is something out there...

    Theo

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Philip Herlihy@21:1/5 to All on Sat Oct 2 12:49:38 2021
    In article <aPf*s6Cvy@news.chiark.greenend.org.uk>, theom+news@chiark.greenend.org.uk says...

    Philip Herlihy <thiswillbounceback@you.com> wrote:
    In article <cPf*1Gxvy@news.chiark.greenend.org.uk>, theom+news@chiark.greenend.org.uk says...

    Thanks. I don't know how old the original PC was - all I have now is the disk
    itself. It's a seagate ST3491A, stamped with a date of January 1995.

    Do you know what OS it was running? That drive is about the borderline between when software and drives switched from CHS (cylinder/head/sector) to LBA (logical block addressing). It is possible current software and
    adapters don't handle CHS correctly, although it may be that the disc is equally readable in either format.

    If the original driver used the CHS in the wrong order you can end up with a permutation of your disc blocks, which makes the drive appear to contain garbage. With LBA there's no ambiguity.

    I was particularly interested in what you said about SMART, which may or not be
    'translated' by adapters. I guess what I'm fumbling to understand is what translation goes on. Clearly, in something trivial like an E-SATA to SATA cable/adapter it's just a case of mapping wires. But with a USB interface there is 'electronics' in there. What's it doing? What can't it do?

    USB Mass Storage is based on the SCSI protocol. IDE and SATA are based on the ATA protocol. It's a bit like English and French, and the adapter has
    to understand both fully to translate between the two. If you don't have
    the vocab to understand what the Frenchman says, you can't say the English equivalent - which is what happens with SMART.

    (There are ways to say 'just quote this French phrase and he'll understand' which cover some of the more obscure corners, but it still needs somebody to speak French)

    The trouble with adapters is the firmware running on them is not always as comprehensive or fully-featured as you might hope, since the chips are being knocked out for $1 a time.

    My concern is the CHS v LBA thing might blindside the USB adapter - while a native Frenchman might understand Norman French, somebody who barely speaks modern French might not.

    Current thinking is to try a PATA to SATA adapter and proceed carefully. I've
    found that if Acronis True Image can 'see' a disk, then an image can subsequently be 'mounted' and the files read - that's all read-only as far as
    the source disk is concerned. I might see if Testdisk can see the disk via such an adapter, but it's possibly more of a gamble to ask Testdisk to 'correct' the partition. I may just end up with my old IDE-capable PC all over
    the kitchen (and the missus all over me...).

    I'm not sure to what extent the cylinder/head/sector registers are
    replicated in SATA, or whether the interface is entirely LBA by that point. But I don't think it would be too dangerous to try - worst case it won't be detected, or appear as garbage.

    I would not ever think of writing to the disk, you can't be sure that won't damage something and it's your only copy. Get a bit image of it that you
    can duplicate, and then work on a copy of that (if you like, write it to a USB stick so it's easy to mount etc. Or connect it to a VM as a virtual disc)

    I keep telling myself I must get on top of Linux. But to operate safely, especially using a 'foreign' filesystem you do have to have a certain minimum
    of contextual understanding or you're going to screw something up. That might
    be relatively quickly achieved, especially for someone with a (historic) background in Unix (SunOS, anyone?) but it'll take more time than I currently
    have. Competent folk often lose sight of the breadth of their knowledge!

    Agreed. I'm not sure what the best user-friendly danger-free recovery toolkit is, perhaps there is something out there...

    Theo

    Thanks for this - I love the analogy with Norman French. As soon as I get a chance I'll give it a go - I found I already have an IDE>SATA adapter and a SATA/E-SATA cable in my bits box. I'll image the disk first. If Testdisk can see 'lost' partitions I'll be tempted to let it fix them, as that's a single write operation, and a calculated gamble. I'm imagining that if the translation is bad, everything will be scrambled so Testdisk won't see anything useful.

    Looking at your headers, do the names Wilkes and Wheeler mean anything to you, Theo?

    --

    Phil, London

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Theo@21:1/5 to Philip Herlihy on Tue Oct 12 21:22:47 2021
    Philip Herlihy <thiswillbounceback@you.com> wrote:
    Looking at your headers, do the names Wilkes and Wheeler mean anything to you, Theo?

    They do indeed :) Maurice was a regular visitor for beer and reminisences about his vacuum tubes. I didn't know David for long before he died but
    ISTR he was rather good at our lunchtime Times crossword...

    Theo

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Philip Herlihy@21:1/5 to All on Wed Oct 13 10:30:29 2021
    In article <dPf*LUxwy@news.chiark.greenend.org.uk>, theom+news@chiark.greenend.org.uk says...

    Philip Herlihy <thiswillbounceback@you.com> wrote:
    Looking at your headers, do the names Wilkes and Wheeler mean anything to you, Theo?

    They do indeed :) Maurice was a regular visitor for beer and reminisences about his vacuum tubes. I didn't know David for long before he died but
    ISTR he was rather good at our lunchtime Times crossword...

    Theo

    David was 'mentor' for my project (Diploma course, scraped**), which was in automatic translation of one computer language to another. He made a comment that really struck me even then (1986) about how we should start looking at "data compilers". How prescient was that?

    **I went to all 17 lectures on Denotational Semantics and to this day I'm not sure what that is.

    For anyone puzzling over these exchanges, Maurice Wilkes and David Wheeler (with Stanley Gill, who know little about) wrote the first ever book on computer programming "The Preparation of Programs for an Electronic Digital Computer" and went on to launch the world's first taught course in that subject. (Which I took.)

    --

    Phil, London

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)