Saab Dastard
26th Mar 2016, 23:17
One of my 3 Win 7 x64 PCs went belly-up on Wednesday morning. I don't know if it was the mobo or the processor, but it didn't post. No problem, the system disk (SSD) and data disk are still fine, I thought, so it should be easy to transfer to a new / upgraded system (plus I have backups).
Time for an upgrade, I thought, so I bought a new mobo, CPU and RAM (and a bigger system SSD) and duly installed them into the case, reusing PSU, graphics & sound cards. So far so good, all booted nicely into the UEFI BIOS. I then decided to clone the system SSD to the new SSD to ensure I always had the original disk in case of problems with the migration from the old AMD Athlon system to the new intel i5 / Z170 mobo.
I thought the quickest way to do the clone would be to put a spare SATA disk (that used to be the system disk of another, similar, AMD Athlon PC that I previously upgraded to SSD) into the newly built PC, then test the migration from AMD to intel on that, then clone the system disk from the dead system to the new SSD. The cloning would then run at SATA 3 speeds (waaaay quicker than USB 2 which was the alternative).
That all worked perfectly, Win 7 recovered beautifully from being transferred from an old AMD system to a brand new intel platform, with different graphics cards, so I had great confidence that the approach was correct, and that it would be a simple matter to drop the newly cloned system disk back into the PC as the system disk and run the repair to get Windows up and running without losing any user settings or having to re-install everything from scratch.
Well, that's not how it turned out. The cloned SSD would not boot into Windows. Neither would the original. Cue much head-scratching and frantic reading about the differences between MBR and GPT disks, and about the newer UEFI BIOS, and playing around with many and various settings therein. But since I had booted successfully off one old MBR hard disk from a similar system, why would this one not do so?
I tried a lot of tricks, involving clean Windows installs on the new disk, then copying over just the boot partition - in several permutations, involving several systems, including my laptop. No joy.
It then dawned on me that the problem was this: The dead system had originally had an IDE system disk, that I had cloned to a SATA SSD, but I had never changed the disk to run in AHCI mode from IDE mode, and the new mobo didn't support SATA in IDE compatible mode, so it was never going to work. The difference with the other PC was that it was SATA AHCI, even though still an MBR disk, which is why it booted on the new system.
I then tried editing the old registry from "outside" to try to fool Windows into thinking that it was in AHCI mode, but - I assume - because there were no AHCI drivers loaded, Windows still couldn't boot. Not having that disk in a bootable system for Windows to find and load the correct AHCI drivers, and not knowing what they were, or where exactly to put them, I finally gave up, 12 hours after starting the rebuild process. And I HATE giving up!
Having thrown in that towel, it took a further day and a half to do a complete re-install of all applications and user preferences on the new system. All is now well (and over 3 times faster, according to Passmark) on the new system.
So the lesson from all this? If you have a SATA HDD in IDE mode, change to AHCI before it's too late! Or more generally, carry out that kind of preventive upgrade to maintain forwards compatibility when it's still possible. Oh, and clone your system disk before doing this kind of migration!
SD
Time for an upgrade, I thought, so I bought a new mobo, CPU and RAM (and a bigger system SSD) and duly installed them into the case, reusing PSU, graphics & sound cards. So far so good, all booted nicely into the UEFI BIOS. I then decided to clone the system SSD to the new SSD to ensure I always had the original disk in case of problems with the migration from the old AMD Athlon system to the new intel i5 / Z170 mobo.
I thought the quickest way to do the clone would be to put a spare SATA disk (that used to be the system disk of another, similar, AMD Athlon PC that I previously upgraded to SSD) into the newly built PC, then test the migration from AMD to intel on that, then clone the system disk from the dead system to the new SSD. The cloning would then run at SATA 3 speeds (waaaay quicker than USB 2 which was the alternative).
That all worked perfectly, Win 7 recovered beautifully from being transferred from an old AMD system to a brand new intel platform, with different graphics cards, so I had great confidence that the approach was correct, and that it would be a simple matter to drop the newly cloned system disk back into the PC as the system disk and run the repair to get Windows up and running without losing any user settings or having to re-install everything from scratch.
Well, that's not how it turned out. The cloned SSD would not boot into Windows. Neither would the original. Cue much head-scratching and frantic reading about the differences between MBR and GPT disks, and about the newer UEFI BIOS, and playing around with many and various settings therein. But since I had booted successfully off one old MBR hard disk from a similar system, why would this one not do so?
I tried a lot of tricks, involving clean Windows installs on the new disk, then copying over just the boot partition - in several permutations, involving several systems, including my laptop. No joy.
It then dawned on me that the problem was this: The dead system had originally had an IDE system disk, that I had cloned to a SATA SSD, but I had never changed the disk to run in AHCI mode from IDE mode, and the new mobo didn't support SATA in IDE compatible mode, so it was never going to work. The difference with the other PC was that it was SATA AHCI, even though still an MBR disk, which is why it booted on the new system.
I then tried editing the old registry from "outside" to try to fool Windows into thinking that it was in AHCI mode, but - I assume - because there were no AHCI drivers loaded, Windows still couldn't boot. Not having that disk in a bootable system for Windows to find and load the correct AHCI drivers, and not knowing what they were, or where exactly to put them, I finally gave up, 12 hours after starting the rebuild process. And I HATE giving up!
Having thrown in that towel, it took a further day and a half to do a complete re-install of all applications and user preferences on the new system. All is now well (and over 3 times faster, according to Passmark) on the new system.
So the lesson from all this? If you have a SATA HDD in IDE mode, change to AHCI before it's too late! Or more generally, carry out that kind of preventive upgrade to maintain forwards compatibility when it's still possible. Oh, and clone your system disk before doing this kind of migration!
SD