56 thoughts on “Computer Problem”

1. This isn’t a problem, it’s an opportunity to get a new machine! 🙂

   1. It’s not that old. It’s a Ryzen 5 3600x.

      1. Um, that’s your problem, right there!

         Disclosure: I am “long” INTC.

         1. Funny, I just made $1200 today being long INTC, in a bull call spread.

2. Hrm… The failure of rescue booting is troubling. Have you tried swapping the primary hard drive with another computer’s HD? If that doesn’t work, I’d say you have a dead motherboard. I just bought one for a Ryzen 5 3600 for about $100.00, though I haven’t swapped it out yet. In my case a USB host controller died and took out half my USB ports.

   1. Well, I want another X570. I’m looking at this one, which seems to be a bargain.

      1. Mine is an ASRock B450M/ac, which is what my computer came with. I stayed with it to avoid any mechanical or cabling issues swapping it out. I would’ve already done so, but with all the political chaos going on, I don’t want to risk any downtime when my machine is still running fine, absent half the USB ports being dead.

3. Rand,

   Haven’t had a problem like the one you describe in a very long time, but…you might consider swapping out the power supply, since they typically have separate (read: fail-able) power leads for different items embedded in or connected to the MB.

   1. Are you saying he needs to try re-routing auxiliary power?

      1. But his motherboard has been sending fake live signals to throw him off track….

   2. Or at least look up your power supply online and see if you can get cable voltages and use a DMM on the leads. Most motherboard cable configs are standardized around a few voltages. Too lazy to look it up, but you can, should be straightforward to find.

4. Do you have a bootable Linux USB flash drive? Have you tried that? Those can come in handy, esp. the ones that allow you to run fsck on your hard drive.

   Could be memory. Memtest will help with that. Or if you don’t have too many DIMMs try removing all but the proper ‘first’ one and then swapping it if still no joy.

   And then there is my all time favorite motherboard trasher. Try replacing the battery on the TOY clock. Known to trigger sad Macs on the ol’ Quadra 605s that the MAC dealers at the time would only be way to happy to sell you a $400 motherboard ($125 from the mail order warehousers) to replace a $12 battery. Let’s hope your nearly new Ryzen 5 3600x doesn’t suffer from that. I don’t think it should.

   1. BTW you are at your quota for the year Rand. Unlike Medicare I have a reverse deductible aka a cap. Will help you again after Jan. 1. 🙂

   2. If the BIOS comes up with the correct date and time and not the one it had when you last shut if off, it’s likely not the TOY clock battery. But I had to throw that out there because, well, because all you MAC bigots… I was once one too…. 🙂

      1. Or a totally bogus date and time. Another tell tale…

         1. Seriously, folks. The BIOS coin battery is something to check — it is cheap enough to just replace.

            I’ll be here all week. Try the veal . . .

   3. I second the recommendation for booting a live Linux thumb drive. I had an update kill grub one time and I fixed it with a boot repair live usb drive. I also recommend having up-to-date kernel versions of Parted Magic and Clonezilla on hand. They’re great for testing what’s alive and what’s not.

      1. +1

5. Rand, my anecdotal experience is that ASRock is a brand to avoid.

   1. I’m looking at an MSI like hers, for thirty bucks more. But that ASRock did have good reviews.

6. Rand, you seem to have an above average amount of computer problems.

   1. That seems unlikely to me. I have an above-average amount of reporting computer problems to this blog.

      1. I, for one, find these computer-problem threads very educational. I’ve learned a lot from them – both the posts and the comments.

         I have no useful knowledge on this particular problem, so all I can say is, best of luck, and please do post an update if/when an answer is found.

7. Since it will POST and you can bring up the config screen, it’s probably a disk problem. Look in the config for the disks and see if they show up. A fix that works at least half the time is to simply disconnect and re-seat ALL of the cables on both ends. You may have to fiddle with the boot order to get it to boot to USB. there are several good diagnostic disk images around and I’m not current enough to make a recommendation. If re-seating the connections don’t fix it, probably a dead or corrupted disk.

   1. I tend to agree with this assessment. Good advice.

      1. If you aren’t using an SSD and it turns out to be a dead disk, now might be the time to consider a switch.

8. RIP Gen Chuck Yeager. 97

   1. Broken ribs, a broom handle, a trusty engineer sidekick and a stick of:

      http://4.bp.blogspot.com/_03nCcx-2AGA/TPlnLfXGGTI/AAAAAAAAAGw/Z3CsxzXJgU4/s1600/Beemans+Gum.jpg

9. Did you try

   Swapping out the sata cables.
   Different sata port.
   Swapping out the power supply.
   Failsafe bios settings.

10. [Tuesday afternoon update] Sounds like a configuration problem post crash that is prohibiting a reboot. Are you running a RAID array? If so make sure you have all your settings set up correctly.

    A good way to check the configuration of your system disk is boot up on your USB flash drive and then mount your system disk so you can inspect all your configuration settings. One possible problem I’ve read about is if your swap partition is mis-configured it can cause a boot hang. If you start a boot and be patient (several minutes) see what kind of messages pop out of the kernel. That would be most helpful.

    1. Also make sure your BIOS settings are correct for your system disk. Are you supposed to be setup for UEFI boot or something else? Boot order in the BIOS correct? Etc.

    2. I’m not running a RAID array. I don’t think it’s a BIOS issue. It has no problem loading grub. But when I select a kernel (or it defaults to boot it), it does the spinning Fedora thing ad infinitum, and doesn’t boot. There are no error messages when I boot from the USB. and I’ve done nothing to the BIOS other than telling it to prioritize USB for booting.

       1. It sounds like a grub configuration problem, or possibly what Grub is pointing to is corrupt. The latter seems unlikely as you have problems with multiple kernels. I suggest trying booting a live boot repair usb and then letting it reinstall grub for you.

11. From a (very) quick review of people having similar issues it sounds like you are missing a BIOS boot partition on your disk. I don’t know how you made your system disk. But you might want to check your partition table to make sure that this partition exists and then re-try the grub2-install. Or that the existing boot partition is marked for BIOS boot. HTH

    1. This is the only boot partition:

       EFI grub2 mach_kernel System

       What should I look for? And what changed? It’s been booting fine ever since I built the machine. I didn’t make any changes to anything Sunday night.

       1. That seems really odd for your configuration. But maybe I’m behind the times. I would have expected your boot partition to say BIOS grub2 linux_kernel System. AFAIK Fedora/RedHat is still Linux which is not mach. Mach is for Gnu Herd. Try some other settings for your boot partition. I have no idea why it is saying this.

          1. Some different switches on grub2 install? I don”t know if it tries to partition a boot partition on your disk as part of what it does, but if it does, it seems the defaults are wrong for Linux.

             1. Speculating that you may have a bad (i.e. failed) sector in your boot partition. But you need to fix the boot partition for BIOS/Linux or perhaps EFI/Linux and try grub install again. What are you using for partition software?

12. See also this page:
    https://www.thegeekdiary.com/centos-rhel-7-how-to-reinstall-grub2-from-rescue-mode/

    You might need the additional commands listed under “Re-installing grub2 on UEFI based machines”.

    1. Go stepwise here. You need to chroot etc.

    2. How do I get into rescue mode? I don’t have that disk. I know where my installation is, but I’m not sure how to mount the sysimage manually.

13. OK, weirder and weirder. I’m following the instructions on this page (just the last one to reinstall grub and shim), but when I do the dnf, it times out getting to the repositories. And I can’t ping Google. Yet I’m able to web surf. Riddle me that, Batman.

    It’s actually very simple. The Internet hates you Rand.
    Try again a little later. Was ‘this page’ supposed to be a link?

14. Try this. Read the WHOLE thing first…

    https://thomas-leister.de/en/repair-fedora-efi-bootloader/

    Basically Rand’s problems are he is using a forever-beta O/S….
    Why dnf would remove key boot files from your system disk without your knowledge (or consent) is beyond me….

    1. That’s similar to the page I forgot to link. But it does me no good if I can’t access the Internet from my terminal.

       1. Did you try the steps to move resolve.conf and replace it with a DNS nameserver of 9.9.9.9 or some other known to work one?

          Also I thought the suggestion of going with hard-wired Internet vs WiFi wise.

          1. The machine is on ethernet, not wifi. I could change the DNS, but that doesn’t explain why Firefox has no problem with DNS, but the shell does.

             1. That is bizarre. Maybe the automatic network config (is this DHCP off an ISP router?) is redirecting you to a bad proxy that only does HTTP?

                Can you do the network settings manually?

15. Fedora, worth every penny of Rand’s time spent working around it.

16. Would anyone happen to know whether Fedora auto-updates?

    The reason I ask is that Rand’s issue reminds me of when a client of mine needed my help to fix a Windows-based machine, which was suddenly unable to correctly see its ethernet. The issue turned out to be a windows update, as he had his machine set to allow updates automatically (never a good idea), and the “update” couldn’t see the on-motherboard ethernet. I took the easy/fast way out (he’s a day-trader, and needed access urgently) and installed an ethernet card.

    So, though it’s a very different OS, my first suspicion would be an update (assuming Fedora does that, and assuming Rand has it set to do it automatically). Just a guess.

    1. Fedora doesn’t auto-update. I do it. I didn’t do an update Sunday night.

       1. You system disk could have been corrupted long before Sunday night. You wouldn’t have noticed it until your system was forced to reboot.

17. If you can’t resolve the boot partition issue you can always just grab a new or spare disk and re-install on it, then once you’re back you can mount the old disk and copy any files you need to preserve off it. This can save a ton of time and frustration. My advice, grab a copy of CentOS 7 or (RH7 w/support) install it and stay away from Fedora.

    1. There is no data on the OS SSD; I keep that on a hard drive that I overmount on /home. I’m just trying to preserve this one because it’s a PITA to reinstall all the packages I use from a new install.

       1. It’s time consuming but shouldn’t be a PITA. How’s your scripting kung fu? You can install your add-on packages from a script you write and maintain on a USB flash drive. After a new install copy it over to your new system disk and run from there.

          It’s all a matter of trade-offs, usually time.

          1. It would have been nice for past me to do that for me, but he didn’t. The bastard.

18. Pingback: COmputer Problems, Continued | Transterrestrial Musings

Comments are closed.