1. Download our Official Android App: Forums for Android!

Windows 10 meltdown -- literally!

Discussion in 'Computers' started by lunatic59, Dec 8, 2019.

  1. lunatic59

    lunatic59 Moderati ergo sum
    Moderator Thread Starter
    Rank:
     #3
    Points:
    4,238
    Posts:
    39,343
    Joined:
    Jun 12, 2010

    Jun 12, 2010
    39,343
    40,219
    4,238
    Male
    IT
    Pennsylvania
    As the end of the year approaches and support for the venerable Windows 7 comes to an end, I am tasked with making sure everyone in our enterprise is using Windows 10 by either upgrade or new machine. My process has been to take a new or upgraded machine, do a complete fresh install of Windows 10 in it's latest iteration (currently 1909), install an office 365 license, malware/anti-virus endpoint protection, a few utilities for remote maintenance and the appropriate shortcuts to our ERP servers, depending on the users need for access. I then restore their files and configs from the previous night's backup. From that point it's a simple 5 minute swap out to get them on the new machine.

    If their old machine is viable to run Windows 10, I will clone the boot disk to an SSD, swap out the hard drive and run the Windows 10 upgrade on the SSD. Once the upgrade succeeds, I verify Windows is activated with a digital license and proceed to completely wipe the SSD and do a fresh install of Windows 10 in preparation for the next swap. The original untouched HDD gets labeled by user and decommission date and put on a shelf in the data center in case anything needs to be retrieved later on.

    I am relating all this to make a point that I am diligent when it comes to user data and uptime.

    So there I sit, watching the Windows update process complete several updates on what should be the final PC in my retiring Windows 7 project and I think that I haven't update feature versions in my own PC in a while. While I do apply all security and bug patches, new features aren't a high priority. I think, "what the heck" let's bring my PC up to 1909 since it was currently only running 1803 and I start the update process. I've done it over and over again with a 100% success rate, so what's the risk? (famous last words).

    When it gets to the first reboot, i get a "no OS found" error. Hmmm. I unplug everything except power, monitor and the boot SSD and reboot. This time it starts the Windows update process but hangs up. No problem, Windows has been pretty reliable at fixing itself in situations like this, even if it has to roll back to previous versions. No soap. Completely borked.

    I end up having to reinstall a fresh version of Windows 10 directly to 1909, reinstall my applications and settings, etc. but when I plug in the 2 TB disk I use for file storage, it can't find the drive, not even in the BIOS. Upon closer examination, I see where my problems started. Disk hardware. Apparently the upgrade process stressed the hardware just enough to generate enough heat to make a connection fail on the ICB. The solder joint on the right failed and in removing the hard disk from the case I pulled off the resitor and part of the copper circuit path. :(

    hhd_parts.jpg hhd_close1.jpg

    While I did have all of the important files backed up, there were some old installations that would surely be a pain in the data bus to track down and recopy. Unfortunately all I had at work was a little battery powered soldering iron that couldn't give me a hot enough or fine enough tip to resolder the component.

    I channeled my inner McGyver and took a spare RJ-45 connector and a heavy rubber band and did the following "fix" to get the disk to spin up long enough for me to get the files off of it. The old mechanical disk is now replaced with a pair of Samsung 1 TB SSD's.

    hhd_fix.jpg
    hhd_fixclose.jpg

    The plastic jack put just enough pressure on the part to hold the connection until I was finished.
     

    Advertisement

    olbriar, Unforgiven, kate and 6 others like this.
  2. James L

    James L Android Expert
    Rank:
     #62
    Points:
    343
    Posts:
    1,511
    Joined:
    Dec 29, 2014

    Dec 29, 2014
    1,511
    873
    343
    Male
    Owner https://rootandroideasy.com/
    East Peoria IL
    Nicely done.
     
    lunatic59 likes this.
  3. rootabaga

    rootabaga Extreme Android User
    Rank:
     #10
    Points:
    2,473
    Posts:
    7,623
    Joined:
    Aug 11, 2014

    Aug 11, 2014
    7,623
    28,010
    2,473
    More or less IT, lots of hand-holding
    Crazyville, CA
    McGuyver’s got nothing on you. Well done.

    We’re in the middle of a W7-W10 project as well, but the guys who are “in charge” of it are more interested in gaming, shopping and making lattes (and excuses!) than upgrading PCs. (I can’t even get them to at least finish the users who are on our legacy file server, which also goes EOS in January.) I guess it’s a good thing I didn’t get promoted to the boss position, because I’d be seriously disliked by this team, and especially their “supervisor,” and yes, I do indeed use that term very loosely.
     
    olbriar, Davdi and lunatic59 like this.
  4. lunatic59

    lunatic59 Moderati ergo sum
    Moderator Thread Starter
    Rank:
     #3
    Points:
    4,238
    Posts:
    39,343
    Joined:
    Jun 12, 2010

    Jun 12, 2010
    39,343
    40,219
    4,238
    Male
    IT
    Pennsylvania
    I hear you there. My wife works for a pretty big enterprise and she's constantly bringing home stories of hoe the "help desk" is anything but helpful. They try to bull#$%^ their way through most of it. Once I met her for lunch and IT support was there "fixing the printer". They were commenting on how this was going to take a while because the printer must have a failed component. It was pretty clear to me that their printer had renewed it's IP using DHCP and the queue was pointing to the wrong IP. Of course when I suggested that, they rolled their eyes and dismissed my suggestion -- at least until we left for lunch. When she got back the printer was miraculously "fixed". I wonder what the problem was? *cough* ;)

    We retired all our Server 2008 machines last year and virtualized our environment. It's so much easier to manage and monitor when the hardware is built like an M1 Abrams and there are only a couple of machines hosting all our servers. And the best part is when we need to add a server or setup a testing environment, it's just a matter of spinning up a new one. :)

    I swapped out the second to last PC this morning and the final one is waiting for next week. I try to do this when the user has a scheduled day off, just to keep impact to a minimum.
     
    rootabaga likes this.
  5. rootabaga

    rootabaga Extreme Android User
    Rank:
     #10
    Points:
    2,473
    Posts:
    7,623
    Joined:
    Aug 11, 2014

    Aug 11, 2014
    7,623
    28,010
    2,473
    More or less IT, lots of hand-holding
    Crazyville, CA
    Good luck with the wrap-up and kudos for minimizing user impact.

    The client support team is actually quite good at resolving trouble tickets. It’s just projects that bog them down, because it’s simply not a priority to them. (They start off great and then it fizzles and moves forward in fits and starts, and to me that falls on the team supervisor since the team follows his lackadaisical leading.) Above us all we only have an interim manager and he’s trying to smooth things over rather than ride herd on projects. Rather than get them to move, we’ve budgeted to pay M$ for updates on the remaining legacy Win7 machines, which means I have to maintain WSUS for them, too. All this crap cascades both up and down, as you well know. :(

    Have a great Christmas, knowing you’re set for 2020. ;)
     
    Davdi, Unforgiven and lunatic59 like this.
  6. lunatic59

    lunatic59 Moderati ergo sum
    Moderator Thread Starter
    Rank:
     #3
    Points:
    4,238
    Posts:
    39,343
    Joined:
    Jun 12, 2010

    Jun 12, 2010
    39,343
    40,219
    4,238
    Male
    IT
    Pennsylvania
    Ugh ... The wheels of industry turn slowly, even when greased with cash.
     
  7. Davdi

    Davdi Android Expert
    Rank:
     #72
    Points:
    273
    Posts:
    1,264
    Joined:
    Jul 4, 2012

    Jul 4, 2012
    1,264
    712
    273
    Semi-retired developer
    The Shire
    I had similar 'fun' a while back- nursing four very out of date servers producing management reports for a large UK retail chain, three of them failed over a two week period, so one was left running 24/7 to just about keep up. I was working 24 hours straight over Sunday/Monday. I asked the ops manager what I should tell the Chairman if/when the last server failed and he didn't get his Sunday trading summary. He just said 'What do you want' so I gave him an over specified wish list - And he agreed! Funny hoe the money was suddenly available.
     
    rootabaga and lunatic59 like this.

Share This Page

Loading...