This video is perfectly applicable, the rot that sets in in a large company when you have no competition to counteract it is exactly what has happened here.

Not sure a short summary will cut it.

They had no competition for a long period and ended up with an accountant CEO that caused their R&D to stagnate massively. They had a ton of struggling and failing to deliver all in most areas, and they wombled about releasing CPU generations with ~4% performance uplifts, probably saving a few bucks in the process.

AMD turned back up again with Ryzen and Epyc models that were pretty good and and an impressive pace of improvement ( like ~14% generational uplifts ) that caused them such a fright that they figured out they had to ditch the accountant.

Pat Gelsinger was asked to step up as CEO and fix that mess. They axed some obvious defective folks in their structure and rushed about to release 12th generation products with decent gains by cranking the power levels of the CPUs to absurd levels, this was risky and it kind of looks like they are being bit with it now.

Server CPU sales are way down because they are just plain uncompetitive. They have missed out on the chunk of money they could have got from the AI bubble because they never had a good GPU architecture they could leverage over to use. They have been shutting down unprofitable and troublesome divisions like the Optane storage and NUC divisions to try and save money, but they are in a bad way.

The class actions mentioned elsewhere in the thread are probably coming because the rush to make incremental improvements to 13th generation and 14th generation CPU’s resulted in issues with power levels and other problems that seem to be causing those CPU’s to crash and sometimes fail altogether.

Yeah, I reckon having a split of the frontend and the backend results in about half the complexity in each. If you have multiple frontends you can upgrade whatever the least important one is to see if there are any problems

I didn’t really answer your original question.

When I was using NUC’s I was using Linux mint which uses cinnamon by default as the window manager. Originally I changed it to use some really minimal window manager like twm, but then at some point it became practical to not use one at all and just run kodi directly on X.

If I was going back to a Linux frontend I’d probably evaluate libreELEC as it has alot of the sharp edges sorted out.

I used to run kodi on linux on intel NUC’s connected to all our TV’s a while ago. I don’t remember it being particularly unreliable. The issue that made me change that setup was hardware decoding support in 4k for newer codecs.

What I’ve had doing that frontend function ( kodi, jellyfin, disney plus, netflix etc ) for the last few years is three Nvidia shield TV pro’s which have been absolutely awesome. They are an old product now and I suspect Nvidia are too busy making money to work on a newer generation version of them,

The biggest surprise improvement was how good it was being able to ( easily ) configure their remotes to generate power on / off and volume up and down IR codes for the TV or the AV amp they were using so you only need a single remote.

Separating the function of the backend out from the frontend in the lounge has reduced the broken mess that happens around OS upgrades drastically.

I replaced mythtv with tvheadend on the backend and kodi on the frontend like 5 or 6 years ago.

The setup and configuration at the time on mythtv was slanted towards old ( obsolete ) analog tuners and static setup and tvheadend was like a breath of fresh air in comparison where you could point it at a DVB mux or two and it would mostly do what you want without having to fight it.

I’m not sure how much longer I’ll want something that can tune DVB-S2 and DVB-T though. Jellyfin and friends handle everything other than legacy TV better than kodi these days.

I got a pretty nice Yamaha bluray player that was an appropriate match to my home theatre amp.

Put a bluray in it, got a piracy warning, a few unskippable ads for other movies, an obnoxious excessively drawn out animated menu screen that stuttered like hell and was laggy to use.

Pulled the bluray back out of it, stuck it back in the DVD drawer and proceeded to download a copy of the movie to watch. Been doing that ever since.

The most impressed I’ve been with hardware encoding and decoding is with the built in graphics on my little NUC.

I’m using a NUC10i5FNH which was only barely able to transcode one vaguely decent bitrate stream in software. It looked like passing the hardware transcoding through to a VM was too messy for me so I decided to reinstall linux straight on the hardware.

The hardware encoding and decoding performance was absolutely amazing. I must have opened up about 20 jellyfin windows that were transcoding before I gave up trying and called it good enough. I only really need about 4 maximum.

The graphics on the 10th generation NUC’s is the same sort of thing that is on the 9th gen and 10th gen desktop cpu’s, so if you have and intel cpu with onboard graphics give it a try.

It’s way less trouble than the last time I built a similar setup with NVidia. I haven’t tried a Radeon card yet, but the jellyfin docs are a bit more negative about AMD.

A couple of seagulls made their nest in the cooling vent for the radiator of one of our backup generators. I caught it on our security cameras and mentioned it to management which resulted in folks being dispatched to evict them and clean up the giant pile of sticks and other junk they had dragged in.

Not sure what would have happened next time the thing started, so it was probably for the best. I still felt bad.

I just read the update to the post saying that the issue has been narrowed down to the NTFS driver. I haven’t used NTFS on linux since the NTFS fuse driver was brand new and still wonky as hell something like 15 years ago, so I don’t know much about it.

However, it sounds like the in kernel driver was still pretty fresh in 5.15, so doing as you have suggested and trying out a 6.5 kernel instead is a pretty good call.

If you haven’t already, try running hdparm on your drive to get an idea of if the drives are at least doing large raw reads straight off the disk at an appropriate performance level.

This is output from the little NUC I’m using right now:

# lsblk
NAME   MAJ:MIN RM   SIZE RO TYPE MOUNTPOINTS
sda      8:0    0 465.8G  0 disk 
├─sda1   8:1    0   512M  0 part /boot/efi
├─sda2   8:2    0 464.3G  0 part /
└─sda3   8:3    0   976M  0 part [SWAP]

# hdparm -i /dev/sda

/dev/sda:

 Model=Samsung SSD 860 EVO 500GB, FwRev=RVT02B6Q, SerialNo=S3YANB0KB24583B
...

# hdparm -t /dev/sda

/dev/sda:
 Timing buffered disk reads: 1526 MB in  3.00 seconds = 508.21 MB/sec

If your results are really poor for this test then it points more at the drive / cable / controller / linux controller driver.

If the results are okay, then the issue is probably something more like a logical partitioning / filesystem driver issue.

I’m not sure what a good benchmark application for Linux that tests the filesystem layer as well is other than bonnie++ which has been around forever. Someone else might have a more current idea of something to use for this.

It might help for the folks here to know which brand and model of SSDs you have, what sort of sata controllers the sata ones are plugged into and what sort of cpu and motherboard the nvme one is connected to.

What I can say is Ubuntu 22.04 doesn’t have some mystery problem with SSDs. I work in a place where we have in the order of 100 Ubuntu 22.04 installs running with SSDs, all either older intel ones or newer samsung ones. They go great.

1988 Nissan Skyline GT with an RB20DET.

It was abandoned by my uncle at our place when he moved overseas and subsequently my sister drove it around a bit. Eventually it leaked coolant from the water pump, overheated and blew a head gasket because she wasn’t paying attention.

I was unemployed and bored and I decided to pull it apart and bought all the bits to fix it. I didn’t really know anything about mechanical stuff at the time, but I am good at logic and try not to be useless at practical stuff even though I’m really a computer geek. I drove it around for a bunch of years after that until I was earning enough money that I could buy something I wanted which was a Mitsubshi EVO 1.

So to answer the question, favorite thing was that I rescued it from oblivion even though I didn’t know much about cars or engines at the time.

This damnable prison of log and ice eats away at my fibre. I find the lack of culture astonishing.

Agreed, it seems like they should have put just a little bit more in the standard feature set so every little window manager doesn’t have to reinvent the wheel.

I learned what a tankie is, which is fun.

I’ve been commenting a bit, whereas on reddit I would only post a comment a few times a year when I could be bothered dealing with the likely burst of negativity that would come as a response to it.

Kind of feels a bit more like Web 1.9 or so from about 2003 which I think was about the sweet spot for minimal rage bait and crazy and still a decent bit of user interaction and scale.

It would be about perfect if you could chop out a few of the folks trying to shoehorn in politics to every little thing.

Appreciate the reply. Which desktop environment are you using?

My only experience with Wayland is also with KDE. Wheres for the 27-ish years before that I’ve used all sorts of stuff with X.

I’ve scripted the machine that drives the frontend for our video surveilance ssytem to place windows exactly where I want them when it comes up.

I use a couple of dbus triggers that make the TV on the wall in my garage go to sleep from the shell, perhaps not tested via ssh though. They were pretty well the functional equivalent of some xset dpms commands that I used to use. Not sure if that is what you were meaning. I think I also had something working that disabled the output altogether. I think that was pretty clunky as it used some sort of screen ID that would occasionally change. Sorry I’m hazy on the details, I’m old.

I’ll try it all out when I get home, I’ve got to find some old serial crap for a coworker in the garage anyway.

Which workflows? Asking because I’d like to experiment with some edge case stuff.

I’m running KDE with wayland on multiple different vintage machines with AMD and intel graphics and it would take alot for me to go back to the depressing old mess that was X.

The biggest improvement in recent times was absolutely pulling out all my Nvidia cards and putting in second hand Radeon cards, but switching to wayland fixed all the dumb interactions between VRR ( and HDR ) capable monitors of mixed refresh rates.

Even the little NUC that drives the three 4k TV’s for the security cameras at work is a little happier with wayland, running for weeks now with hardware decoding, rather than X crashing pretty well every few days.

Last week I did an install of Debian 12 on a little NUC7CJYH to use for web browsing and ssh sessions into work and ended up with wayland by default. Seems to work great.

From what I have experienced, it goes great with intel integrated graphics, great with a radeon card and can be made to work with Nvidia if you are lucky or up for a fight.

We did an address check when we could first order it and about a third of the folks in the office could get it about a year and a half ago. I know the majority of the address checks that we do for commercial locations in tenders come up positive now.

It is not cheap to get an off the shelf router that does a solid job of forwarding multiple gigabits and the vast majority of folks ( me included ) probably will rarely notice the difference outside of speed tests. The last firewall build that I did for home was with a pair of virtual Linux boxes with 10G interfaces just so I could do a 2G or 4G GPON upgrade later on without having to throw everything out.

In New Zealand it seems like 10G GPON services are mostly cannibalizing high quality lit ethernet services at 1G and 10G subrate rather than replacing consumer tails. So more likely a business is going from spending $1500 a month on uncontended 1G to spending $400 a month on contended 4G, rather than a residential user going from spending $150 on 1000/500 to $280 on 2000/2000.

My day job is building ISP networks. It’s been about 20 years since I had a home connection that I didn’t configure up both ends of myself.

I’ve got a 1G / 500M tail into home where I am right now, not that that is particularly impressive. One of the jobs I’ve been putting off at work is standardising our usage of the 10G GPON platform available here in NZ, when I do that I’ll get one of the >1G tails to use at home.

Usually the answer is how ever much I can be bothered building, but my usage is pretty low.