MBR2GPT, Clonezilla, Integration

  • We found new enough Windows installation CD which did include the MBR2GPT tool. Which worked fine after all. It was also requird to enable in BIOS "Windows Boot Manager" boot option. But as far as I know, that option was only available since running the MBR2GPT tool, which created the required GPT partition information for boot. It seems that the bcdboot tool didn't make all the required modifications. Maybe I should have set manually the GPT partition UID to UEFI boot ID or something like that. Anyway, the MBR2GPT was very easy and safe way to make the conversion. After that there was a problem, because recovery partition was blocking extending the system partition. Well, Windows Partition Manager didn't allow us to touch the recovery partition. But gparted moved it beautifully. Interesting thing was that when gparted started to resize the operating system partition, it crashed. Yet that didn't matter much. After booting back to Windows, it was now possible to extended the primary system partition to full disk size, because the recovery partition was moved to the end of the drive with gparted. - What a job, but now it's done, as it was supposed. Just took a little more effort than was initially expected.
  • Next, trouble shooting Clonezilla / partclone issues with Windows 10 drive imaging. More fun, every day. I compiled latest partclone on 64 bit Ubuntu 17.10. Then tested imaging each partition one at time, everything worked. Except of course the MSR partition, which I imaged using dd instead of the partclone. No problems were found. Now there are two remaining issues, either it's how Clonezilla operates with the cloning tools. Or maybe it has something to do with our disk cloning network. Because the error message split - shows is i/o related. This is something which was left open for now, and testing must continue bit later. I'm confident that we'll figure out what the actual problem is, it just takes bit of straddling. Interestingly the problem exists with specific type of computers, but it might be so that as example those compress data faster than other computers and therefore trigger some network specific issue. - Just guessing here.
  • More completely broken websites, now AliExpress has broken their login page so badly, that you can't even login. This is just wonderful trend on many sites.
  • IT project truths, once again got question how much does integration to X cost. Well, well know the cost and requirements when the integration is ready and has been in production for 6 to 12 months and everyone is happy. At that point, if you want specification document, I can write it and tell how much the project costs. And guesses before that, or writing specification (which nobody is actually usually and very unfortunately interested about) is just one way to totally fail. - This is the proven method I've used successfully time after time. - This is also the method which guarantees that project is successful, unless customer pulls the plug due to time / cost constraints. - No matter how many pointless project management or software development books your read, this fact is always the same. And also applies to everything else than software projects too.
  • Air Crash Investigation aka Mayday - Reason for failure: Not following standard procedures and best practices. Nothing new afaik. I'm sure this is happening in every industry over and over again.
  • When doing stuff, there's a lot of things which can be learned simultaneously. So the process was good learning after all. I also learned new information about the Windows OEM recovery partition: The OEM system recovery partition is usually located either at end of disk (for hard disks with capacity below 1 TB) - So that's why I've seen it in several different locations, and that's why the server had it in the way of extending the primary partition, because the old system disk was 500 gigs only and the partition was located at the end of disk space.