Re: domU memory exceeded =?=> spontaneous reboots

[Elliott Mitchell <ehem+xen@xxxxxxx>]

On Sun, Dec 08, 2024 at 02:03:58AM +0000, Mike wrote:
> Elliott Mitchell wrote:
> 
> > Did you in fact look at the output of `xl info`
> 
> Yes.
> 
> > Are you instead guessing based on what you think your
> > computer's setup is?
> 
> "313" is a guess based on what I recall changing, since I don't have that
> configuration anymore.
> 
> > > dom0_mem=4G;max:4G dom0_max_vcpus=4 dom0_vcpus_pin 
> > > xpti=dom0=false,domu=true no-real-mode edd=off
> > 
> > Okay, but 4GB for domain 0 strikes me as an odd value.  If this computer
> > is primarily a server then 1GB should be plenty.
> 
> OK.  But that's scaled back from the default, all 64GB of the memory and
> ballooning.  I can afford it for now.  I can also easily add 64GB to the
> machine.

Okay.  Mostly I was asking questions aimed at ruling out memory
exhaustion as a possibility.  Your numbers are a bit inconsistent, but
with those sorts of numbers that seems unlikely.

> > Is
> > this computer primarily your desktop and you're using Xen to generate VMs
> > to experiment with Kubernetes?
> 
> It's a personal server, no GUI.  I'm migrating my services to Kubernetes.

Okay, 4GB during setup.  For a pure VM server exhausting 4GB is quite
difficult for the control domain.

> > the moment you have a VM machine, it will start eating your network.
> 
> I don't expect much in that regard.  I'll probably have a few dozen
> Kubernetes pods (VMs), and Calico (a Kubernetes network plugin) handles
> all the inter-node (inter-domU) traffic within the domU.  But there will
> be communicaton between every pod and the control node (the other domU),
> and that would pass through the xenbr0.
> 
> 
> FYI, I've been using Xen for, idk, 15-20 years?  It's mostly been set-it-
> and-forget-it.

Okay, the easier to diagnose issues seem unlikely at this point.
Unfortunately this means figuring this out is into hard stages.

In order to eliminate software (Xen) bugs, what is really needed is to
break out a null modem cable.  Then hook up another computer or something
else capable of functioning as a dumb terminal (ha! old tech to the
rescue).  Serial ports are rather more effective for observing Xen
panicing.

Any possibility the computer's power supply is near its limit?  What
could explain the sudden reset is processor usage goes up and pushes
something power related beyond its limit.

For the heck of it, you might try adding "cpuidle" to Xen's command-line.
Then ensure the kernel module "xen_acpi_processor" is loaded.  If this
helps then it could be power or perhaps processor heatsink.


-- 
(\___(\___(\______          --=> 8-) EHM <=--          ______/)___/)___/)
 \BS (    |         ehem+sigmsg@xxxxxxx  PGP 87145445         |    )   /
  \_CS\   |  _____  -O #include <stddisclaimer.h> O-   _____  |   /  _/
8A19\___\_|_/58D2 7E3D DDF4 7BA6 <-PGP-> 41D1 B375 37D0 8714\_|_/___/5445