It’s a NVMe drive in a USB enclosure. Said it was busy or something, if you type eject on the entire drive, it says basically the same thing.
That is when booted into openSUSE that is installed, I even logged out, and went to a terminal with ctrl + alt + f1. No idea what was using it.
So I gave up and booted the openSUSE install ISO. I was able to run fsck in rescue mode doing that.
I tried booting my Windows VM to update it, was going to try using WoeUSB to write it. To see if it can, after converting the VM disk image to a RAW disk image. But if it can’t, write the ISO, then dd to the other partition.
The USB NVMe drive, only has one partition. Perhaps the replacement enclosure is worse. The VM just got an error, no freezing KDE though this time. I had to click the reset button, as VirtualBox froze everything, and I went into a terminal, typed ls on the mount, and it hanged. Couldn’t kill that process, so just typed sudo reboot, but was taking way too long, so gave up and clicked the button.
And perhaps disabling XHCI handoff was a bad idea. Or the Kingston drive in it, is dead already. Nice, I lost my shitty Windows VM, well not completely, it’s also on a USB drive. A SATA SSD in a more reliable enclosure.
Should have tested VM when I got the replacement, apparently that’ll test the NVMe enclosure. And cause it to crap out.
unable to enumerate USB device
Might be a problem.
I/O error while writing superblock
device not accepting address 13, error -71
So Linux doesn’t support USB 3.0?
Perhaps a non rolling distro would be better. I’d think the autosuspend bug would be fixed by now, perhaps not. Let’s see if booting the VM kills it again. Then I’ll try disabling autosuspend, don’t care about that anyways.
Hmm, so XHCI handoff, may have lead me to fix the enclosure, I was having some problem, but only appeared a month or longer after getting the first one. Can’t recall what the problem was, perhaps it wasn’t mounted anymore, so disconnecting.
Perhaps this Plugable NVMe enclosure doesn’t like autosuspend. And perhaps disabling useless power saving, at least useless on a desktop, will let me reboot without having to unmount and unplug and plug back in, to get it connected at the right USB speed. Just run a benchmark after rebooting, and sometimes, it is less then it should be.
The autosupsend disabling instructions can be found here.
It’s been an issue since kernel 2.6.27. Nice. Still waiting for drive to crap out again. Might not.
Just like Proton was updated today, and Just Cause 3 still doesn’t work on Steam Deck, probably works fine on desktop, unless I got temporally banned because of the DRM. If you open too many times, you’ll get temporally banned, and can’t open on anything. If you use a crack, on your desktop, it might work, but crack doesn’t fix it on Steam Deck. I refuse to use Steam Deck until Just Cause 3 works. Cause I want to play Just Cause 3 on it. I could change the Linux distro, that might fix it. Some people say it works, no idea how it works for some people but not others. They never updated the OS? Perhaps.
Oh and Steam might say the game works on the Steam Deck, so they lie. It’s already been reported to them, it just takes forever to get anything fixed, including regressions. Windows Update 100%, and drive hasn’t crapped out yet. I think I have to reboot and then launch the VM. But I think I’ll do the permanent fix, and then reboot. I don’t need autosuspend.
sudo: mkinitrd: command not found
Lies, guess those instructions aren’t for openSUSE. And why does using XHCI handoff fix that problem? Wasn’t a problem until I disabled that. But I don’t think I need to run that command anyways.
Going to re enable XHCI handoff. Windows is installing update now. I’m also going to disable autosuspend for USB, in case that’s causing weird issues. Like around 400 MB/s read on a NVMe drive, that should get over 900 MB/s on whatever USB version that C port is.
Good idea, add usbcore.autosuspend=-1 to my kernel command instead, in case you need some mkinitrd command.
And couldn’t even boot, the NVMe drive wasn’t working, unplugged it, plugged back in, and rebooted, and booted fine. I could have tried clicking ctrl + d or whatever it said. So USB enclosures can’t fix themselves by rebooting the computer? You have to unplug it and back in? No wonder USB isn’t good for storage.
After editing /etc/default/grub, run sudo grub2-mkconfig -o /boot/grub2/grub.cfg.
Now to boot Windows VM, and leave it running for a while, to see if drive craps out a third time. If so, the new enclosure is worse, or the drive has issues. Or both. It could be overheating as well. Going to watch the temperature while the VM is running.
If you want to resell your hosting, or offer free hosting, use FraudLabs, the free plan. Make sure you deny everybody not using their home connection, or maybe a mobile connection. Proxy and VPNs etc, should be blocked from ordering.
Anybody want free 1 MB of space hosting? No bandwidth limit, but you’ll have plenty of space. Enough for a one page website. No support will be provided, and I may stop offering it at anytime. If you use the space to host a PHP file to inject on sites, you’ll be banned. And reported. You won’t get any email. Too bad I can’t disable email on the account from PHP. Perhaps no PHP either. Just SFTP. No insecure FTP.
I’d use the free plan for selling hosting too, I’d just limit it to 500 queries a month. And there’d be a low limit on how many plans I’d sell, only 20 GB of space, and I’m using some. Also, I only need to sell four, to not have to pay my hosting anymore, at least BuyShared.
Good news, the NVMe enclosure, might be stable after disabling autosupsend and reenabling XHCI handoff. Need a motherboard that isn’t anti Linux, MSI hates Linux. The updates installed, and I even ran a benchmark in the VM.
Now to convert the disk image to a RAW disk image.
You can report the autosupsend bug, I’m too lazy and stupid to do so. And it might already be reported. It should be disabled by default though.
How (Not?) to write a RAW Windows 10 VM disk image to a USB drive using WoeUSB:
sudo losetup –find –show –partscan –read-only ‘/media/fast_stuff/VirtualBox/Windows 10/Windows 10.img’
sudo woeusb –device /dev/loop10p2 /dev/sdj
Until it says:
/usr/sbin/woeusb: line 1649: ./Program Files/WindowsApps/Microsoft.UI.Xaml.2.4_2.42007.9001.0_x86__8wekyb3d8bbwe/AppxBlockMap.xml: No such file or directory
So perhaps you have to manually copy the files to the partition it makes. Nice.
rsync didn’t work either. Is there a way to skip broken files? Eh, just use the WoeUSB –partition setting, after using dd of the entire image, and delete the other partitions.
W95 FAT32 that could be the problem.
sudo woeusb –target-filesystem ntfs –device /dev/loop10p2 /dev/sdj
Perhaps that’ll do it. Looks like that might be working. Over 30% done, nice.
And once done running WoeUSB, run losetup -d /dev/loop10, change loop10 to whatever the correct value is.
And looks like it worked, nice.
Going to try booting it. Then I’ll upload this post. And it worked, because I used the ISO, time to try the loop instead, like I meant to do. No I don’t want to install Windows. Just boot my VM on bare metal, in case I need to update my controller’s firmware again. Cause they are dicks, and only have Windows software.
WoeUSB still fails, probably only supports installation ISOs.
But try running sudo dd if=/dev/loop10p2 of=/dev/sdj1 status=progress after it fails. Change the partitions of course, for all the above commands. Perhaps the way Linux is mounting it, it can’t access some files or something, dd might work in that case.
I should dd the partition with the bootloader, so you don’t need WoeUSB. But I’m lazy, so I’ll pass. And you might also need /dev/loop10p1. If it won’t boot. Actually, backup the second partition on USB drive, with the UEFI crap, and dd /dev/loop10p1 to it, then copy the stuff you backed up to it, or the other way. But it’s only 1M. So I need gparted to make it bigger.
It might be easier to install Windows to the drive, bare metal or something, and then dd the big partition.
Or lookup how to convert non UEFI to UEFI.
Perhaps it can boot without that other partition. If dd ever finishes.
And it won’t boot, I got the installer to boot, with all the data still on it. But it looks like the drive is in MBR, and Windows wants GPT. Going to bed soon though, so I’ll change the drive to GPT tomorrow if I feel like it. And lookup how to convert it to UEFI. I mounted the Windows 10 installer ISO, and copied it to the big partition, but all that did was boot the installer, even with that UEFI NTFS bootloader. Which I don’t need anyways. My motherboard supports booting NTFS directly. As there’s two boot options after copying the ISO contents to the drive.
If you can figure out how to install the UEFI bootloader, or get the files, it’ll probably boot on my motherboard, perhaps I need to do it in VM, and convert it to a RAW image again. There’s some bootloader thing for Windows, perhaps it can get the UEFI stuff.