Willing to upgrade - recommended hardware?

[justinclift]

Weird. Well, at least you figured out a workaround and got things working. Sounds like a win.

Hopefully the problem does turn out to just be kernel related, with the USB dongles being bulletproof from here on.

 

[Marc Finns]

Weird. Well, at least you figured out a workaround and got things working. Sounds like a win.

Hopefully the problem does turn out to just be kernel related, with the USB dongles being bulletproof from here on.

Indeed.... but I like to understand why..!

 

[gfngfn256]

Kernel 6.8 has had more than its fair share of problems - you are probably one of these.

 

[Marc Finns]

UPDATE 2 - Unfortunately, the same random freezes happen also with kernel 6.5.
Any other theory? I would hate going back to esxi!
thanks again!

 

[justinclift]

Two possibilities immediately spring to mind. Is the irqbalance package installed, and the service running?

From rough memory it's not installed by default as not every installation needs it. But it's probably worth adding to see if it fixes the problem.

If that doesn't work (and it's unlikely to hurt anything), I guess the next thing to try is seeing if there's info available about what's causing the pauses. It may turn out to be something (other than irqs) having an issue which might be resolvable using a bit of configuration.

With these freezes, are they a permanent hang needing a reboot, or are they more a temporary lock up that comes good after a bit, or something else?

 

[Marc Finns]

Two possibilities immediately spring to mind. Is the irqbalance package installed, and the service running?

From rough memory it's not installed by default as not every installation needs it. But it's probably worth adding to see if it fixes the problem.

If that doesn't work (and it's unlikely to hurt anything), I guess the next thing to try is seeing if there's info available about what's causing the pauses. It may turn out to be something (other than irqs) having an issue which might be resolvable using a bit of configuration.

With these freezes, are they a permanent hang needing a reboot, or are they more a temporary lock up that comes good after a bit, or something else?

The freezes are permanent hangs needing a reboot. No log entries. Consistent with the issue reported for kernel 6.8, except that it happens also with 6.5.

 

[justinclift]

Damn. That can mean any relevant kernel messages aren't getting saved to disk, which makes things harder.

It's kind of a long shot, but I'd _probably_ look at how easy it is to set up a remote syslog machine for receiving kernel messages over the network.

That's not foolproof either, but at least gives some possibility that if there are relevant kernel messages being generated you might be able to see them then figure out WTF is going on.

 

[Marc Finns]

Damn. That can mean any relevant kernel messages aren't getting saved to disk, which makes things harder.

It's kind of a long shot, but I'd _probably_ look at how easy it is to set up a remote syslog machine for receiving kernel messages over the network.

That's not foolproof either, but at least gives some possibility that if there are relevant kernel messages being generated you might be able to see them then figure out WTF is going on.

I could try but I don’t see why it would give additional info compared to the local journal.
The kernel hangs with no logs are widely reported, and I agree it is a significant issue. Frankly, I can’t rely on this thing. If it hangs while away, I would have issues even reentering the house..

 

[justinclift]

Yeah. Unreliability is a killer for me when considering solutions too.

 

[NullNill]

Hello, have you found a solution to this problem? I seem to be running into it as well. I've tried downgrading to kernel 6.8.4-2 but so far the problem persists. I've tried ethtool -K eno1 tso off and it seems like it's working. However it's not a permanent solution as it turns back on whenever I reboot and having to run the command again every reboot is impractical.

 

[Marc Finns]

Hello, have you found a solution to this problem? I seem to be running into it as well. I've tried downgrading to kernel 6.8.4-2 but so far the problem persists. I've tried ethtool -K eno1 tso off and it seems like it's working. However it's not a permanent solution as it turns back on whenever I reboot and having to run the command again every reboot is impractical.

No, I tried everything, but reliability is key in this specific application. I went back to esxi

 

[Retro1982]

Hi,

Just wondering if you did test the ram? I am also using the NUC8i7BEH now for years with ProxMox and some heavy workloads and had no problems so far …

 

[Marc Finns]

Hi,

Just wondering if you did test the ram? I am also using the NUC8i7BEH now for years with ProxMox and some heavy workloads and had no problems so far …

I doubt it is a hw issue. The exact same machine runs literally for years with the same heavy workload on esxi…

 

[gfngfn256]

having to run the command again every reboot is impractical

AFAIK, you can add that ethtool tso off command to the /etc/network/interfaces as a pre-up/post-up command. Search online for the correct procedure.

 

[justinclift]

Yeah. There's going to be some way to have that setting be on or off (as needed) on a permanent basis so it doesn't need to be run every time.

Doing that sounds like it'll fix the problem.