Ask Your Question

xenithorb's profile - activity

2016-08-18 04:56:55 -0600 commented answer How to enable zram for Fedora 22

It's missing an [Install] section in the service, maybe because it was only intended to be used upon installation?

See: https://gist.github.com/danilkuznetso...

2016-04-29 14:13:35 -0600 received badge  Necromancer (source)
2016-04-29 01:03:01 -0600 received badge  Editor (source)
2016-04-29 00:56:24 -0600 answered a question Kernel update 3.12.10 breaks boot: unable to mount /boot/efi - vfat not recognised

Hi there. I'm answering this for posterity because I recently came across this question when I was searching for the very same problem.

So here's what's going on... For whatever reason modules aren't being loaded properly for your rescue kernel, and, while you may be able to boot no problem, and network ( dhclient <dev> ) - what you will most likely not be able to do is use DNF to install the kernel modules you need (my rescue kernel was 4.4.6 for example, and all that was left in the repository was 4.4.8)

The main problem is that Fedora isn't building with vfat inside the kernel - so what happens is even when you go to update your grub.cfg while the rescue kernel is booted. /boot/efi/ is nothing more than a folder on /boot at that point. Recall that on a UEFI system there is; /boot/ and /boot/efi mountpoints. By default /dev/sda1 -> /boot/efi and /dev/sda2 -> /boot So we can regenerate initramfs/grub.cfg and this isn't going to do anything since the partition is unmounted, yet we can not mount the partition.

Bit of conundrum, eh? Well what we do have is 1. A running kernel 2. Presumably network ability 3. Possibly kernel modules for a newer kernel. So let's make sure we have the latest kernel with dnf install --best kernel

Observe now that there should be a /lib/modules/<kernel_ver> corresponding to the version you just installed (or already had installed) after installing the kernel-modules package, which kernel should have pulled in.

So next we have to think.... What do we have at our disposal? Initramfs isn't going to help, that resides on /boot/, adding rd.break isn't going to help, because that stops the initramfs from switching roots - but the only kernel we can boot is the rescue kernel and it still doesn't have the right vfat modules!

Ah! What we do have is our disposal is a working entry in the grub.cfg that we can not reach, and this points to a certain kernel, the rescue kernel itself.

All you have to do at this point is cd /boot and copy the rescue kernel somewhere (preferably by adding a .old to the end or something) and then copy the current kernel that you CAN download stuff for, as the exact name of the rescue kernel.

Then when you reboot and select the rescue kernel again, it will boot the current kernel, equipped with all the necessary trimmings and modules. From here you can get into /boot and /boot/efi and fix things up to your liking, generate initramfs with dracut, etc.

:)

2015-11-21 00:18:26 -0600 received badge  Nice Answer (source)
2015-11-20 11:42:28 -0600 received badge  Nice Answer (source)
2015-11-19 17:47:12 -0600 received badge  Necromancer (source)
2015-11-19 14:08:02 -0600 answered a question Yum autoremove deleting also used packages

Anything installed with PackageKit will be removed by dnf autoremove because they are marked incorrectly and dnf doesn't understand that they are installonly

Solution: don't use the tray updater, and don't use apper / packagekit.

2015-11-10 04:54:23 -0600 received badge  Good Answer (source)
2015-11-09 11:03:32 -0600 received badge  Supporter (source)
2015-11-09 10:14:07 -0600 received badge  Nice Answer (source)
2015-11-09 04:13:08 -0600 received badge  Necromancer (source)
2015-11-09 04:13:08 -0600 received badge  Teacher (source)
2015-11-09 02:45:14 -0600 answered a question Why don't people use their votes in here?

I think psychologically the system here prevents voting. Long time Fedora user, new ask.fedora member. Some posts require >100 somethings to vote, some only require >10. Trying to vote and being rejected with a big red dialog isn't rewarding and will probably prevent me from trying again for quite some time. It's an option I probably won't try again because it isn't clear when you actually have the option to exercise that right. As such, it will be a feature that I simply discover later some day when I think I have the ability. If that intuition is too soon, I will perhaps push trying to vote again even further.

I hope you can see how the policy around voting is what's causing the issue.

2015-11-09 02:45:14 -0600 answered a question Cursor freezes in KDE

https://ask.fedoraproject.org/en/ques...

Please see the answer provided here. What you are most likely seeing with KDE are issues with the nouveau driver. For Fedora 23, you'll either need to attempt to follow the above instructions or downgrade Xorg + install RPMFusion + install akmod-nvidia

2015-11-09 02:45:13 -0600 answered a question fedora 23 freezes constantly

You were most likely using the nvidia drive beforehand and are now using nouveau since the upgrade. I say this only because I've been in the same situation with 22 and it didn't work either and hung all the time, and I happen to know that the proprietary driver isn't yet supporting F23.

Nouveau is notoriously unstable with GK106 and I have been observing these same instabilities. The best I can do for you right now it to suggest that you extract the PGRAPH firmware from the proprietary driver because that blob happens to be more stable.

I'm tracking (hopefully) these issues here: https://bugs.freedesktop.org/show_bug...

You need to follow the instructions on this page:

$ mkdir /tmp/nouveau
$ cd /tmp/nouveau
$ wget https://raw.github.com/imirkin/re-vp2/master/extract_firmware.py
$ wget http://us.download.nvidia.com/XFree86/Linux-x86/325.15/NVIDIA-Linux-x86-325.15.run
$ sh NVIDIA-Linux-x86-325.15.run --extract-only
$ python2 extract_firmware.py  # this script is for python 2 only
# mkdir /lib/firmware/nouveau
# cp -d nv* vuc-* /lib/firmware/nouveau/

What is not included there are Fedora specific instructions, and info about getting PGRAPH firmware loaded. So, what you need to do is this:

$ # This is the instruction that tells the kernel mod to look for the proprietary binary blob.
$ echo "options nouveau config=NvGrUseFW=1" | sudo tee -a /etc/modprobe.d/nouveau.conf 
$ # Make sure you have already done the firmware extraction above at this point. It does expect /lib/firmware/nouveau/
$ # Dracut (initramfs) needs to know about the new firmware
$ echo 'install_optional_items+="$(find /lib/firmware/{nouveau,nvidia} -printf "%p ")"' | sudo tee /etc/dracut.conf.d/nouveau_firmware.conf
$ sudo dracut -f # Let's rebuild the initramfs to include our modprobe.d changes and firmware
$ reboot

This won't give you perfect stability, but will cut down on crashes significantly. I can sometime go double-digit hours without hanging. I think most of the issues that remain are due to some aggressive behavior in kwin / plasma-shell