[DONE (again)] Ryzen vs. U̶b̶u̶n̶t̶u̶ F̶E̶D̶O̶R̶A̶ Ubuntu

main thing to work through right now is disks.

fedora uses LVM, which I am not terribly familiar/comfortable with. I've been trying to add a few disks, but everything I can find is focused on adding drives to a volume: this is exactly not what I want. I want the new drives to have specific paths, so they're clearly segregated from the boot drive. This is both for performance reasons (i.e., these SSDs are much slower than my nvme drive, so I want to keep certain files in certain places), and also because I need to be able to just unmount and physically remove the drives and take them out with me, from time to time.

normally, I'd just add entries for them in fstab. but this seems to cause breakage with LVM.

: (

Why did you choose LVM, I just use 'Standard Partition' for configuring my Hard disk. You can make those changes if you choose Configure my Own disk before install. Im on F25. I know there is a way around your issue though

everyone said it would be easier?
TBH I wasn't really focused on that, at the time. I just let Fedora install, mostly with whatever defaults it wanted. I may re-install, but that's probably not going to be for a while; I want to get things ironed out first. I want to use this machine. :slight_smile:

This is nutso. Just having disks connected at boot is breaking things. If i disconnect them (e.g., unplug the sata power cable) everything's fine. If any or all are connected, I get the fedora logo, but it eventually hangs or drops me into emergency mode.

: (

Maybe I'll have to try re-installing sooner than I thought.

down the rabbithole I go again... although I havent used fedora recently as I'm on a 2011 macbook and all kernels that are available to me on it dont work with wifi. I'll give this a go and hope for the best.

When I first installed f25 with kernel 4.8 it was a royal mess with AMDGPU. But a sudo dnf update got it in a good good space with the kernel 4.10 versions and I'm still fedora myself.

Ryzen well asus motherboards. I been killing them.. No idea why.

1 Like

Yeah, I've been keeping up with updates. This morning brought a kernel update, which I will be installing shortly.

Also, ooOOhh fingers crossed

AMI BIOS 7A34vA4
- Improved memory compatibility.
- Fixed PCIe Hot-plug function issue.

so, SOMEHOW i ended up with the v4 bios (+1), Fedora 25 (-1), and kernel 4.12 (+1).

Also figured out the disks problem, I think. I kept Fedora 26 on the nvme drive, and installed 25 on a regular SSD. In the attempt, the F25 installer kept crashing, and eventually told me it was because, among the disks I had connected, there were * two * LVM's with the same name ("Fedora" - left over from previous install attempts). Wiped those, and the syncing error seems to have gone away.

This is feeling a lot more stable. Over the past couple days, doing actual work on this machine, things had been shaky, with occasional program crashes/misbehavior, one actual lockup, and a surprise reboot. I'm also putting a lot less effort into installing packages, dependencies, and the whole "just working" end of things.

So I'll go a few days with this, and see how it works out.

2 Likes

Are you still going with LVM? Is it important to do so? I've always manual config'd my Fedora installs from anaconda (the installer) Luks partitions /ext4 /xfs and even dual boots. Maybe have a look if you can...

1 Like

Tried overclocking again back to 3.85 after F26?

1 Like

nope; this was another motivator for trying a new install. plain old partitioning this time.

which means I have an appropriately-sized swap partition now, but hibernation doesn't seem to be working anyway. has troubles resuming. it either reboots itself after resuming, or resumes to an empty desktop and just sits there.

I had it back with F26, but at some point during this upgrade it got cleared out. I haven't gone back to re-apply it yet, but I don't see why there would be any problems.

I have noticed, however, that my system is running quite a bit hotter since I upgraded to the latest bios. it idles in the low 50's now, and that's with the case fans turned on (previously, I was only running the cpu and exhaust fans). :frowning:

Hibernation? Your going to power settings and putting a timer and it wont wake from that?

https://browser.geekbench.com/v4/cpu/2994960

of course, I don't know what any of these numbers actually mean.

no, as in,
- hibernate
- unplug
- watch Star Trek
- plug in
- turn on

it comes back but not well. everything is still there, but then it either goes to the lock screen (with no sign-in) and just sits there, or blacks out and reboots.

1 Like

As you've probably seen/read, I'm using the stock heatsink right now, because my hyper 212 turned out to be just slightly too tall atop the AM4 socket. I'm now considering what I might (eventually) get to replace it. Right now the Cryorig H7 is at the top of the list, but I'm looking for other options too.

Soliciting suggestions!

requirements:

  • 155mm or less from bottom of block to top of tallest component (this includes mounted fan position).
  • Very large cooling volume. I want very chilly; also I like massive-looking finstacks.
  • AM4-ready is a big plus; but (of course) adapter kits are ok.
  • Preferably silver and black in color.

Thanks.

arctic freezer 33 is probably your best option. Be Quiet! Pure Rock Slim will also work.
Scythe Mugen 5 PCGH should also fit, but is imo overkill, and more expensive (and weighs 1kg).

1 Like

good options, especially the scythe. and I found the arctic xtreme r.2, which looks like an excellent choice — except for the gaudy side/top panels. anyone have one? know if they're easily removable?

Going by this review, this one isn't particularly good, though. http://tech-reviews.co.uk/reviews/arctic-cooling-freezer-xtreme-rev-2/6/ (it's compared with a mugen 3 generations older than the one I linked, which saw a big performance improvement itself going from the 3 to the 4, and an incremental one from 4->5).

1 Like

I've been seeing this from time to time, on boot. I have to power cycle to continue. In all cases, this only happens when power has been off for a few hours or more, and power cycling always results in a successful boot. Any input would be appreciated.

the first four lines appear on every boot, successful or not. I'm not sure where to start troubleshooting for the following stuff. Thanks!

think the uefi db list has something to do with secure boot/tpm (disable in bios?)
wrt amd Vi: did you enable iommu and svm in bios?

no idea what the other stuff is about / indicative of, sorry. :slight_smile:

1 Like

So, major events / changes:

  • RMA’d the CPU.
  • RMA’d the motherboard.
  • Switched out the 212 evo for a Cryorig C1.
  • RMA’d the motherboard again.
  • Bought a new motherboard (AsRock X370 Taichi).
  • Switched back to Ubuntu, but 17.04 this time (kernel 4.10) and Gnome.
  • Switched GTX 1050 ti for an R7870 (nvidia drivers just aren’t working out).
  • Switched R7870 for an XFX RX 480 8GB.

! ! YAY ! !

Thanks to the amazing backplate, was able to neatly tie down the power cables.
Only downside of this GPU is that the backplate and cooler almost completely cover the PCIe slot latch, so removing the card is very difficult. I have to push the latch down with a spludger, and it’s tense because it’s right above the nvme drive. Don’t slip!

~ ~ AMAYZING ~ ~

Installation, with full-disk encryption, worked first time. Sleep and Hibernation work flawlessly! Still installing stuff, but it’s looking pretty damn good so far.