On Wed, 14 Sep 2022 12:37:41 -0500, Hector wrote in message
<36d765d0-e9a7-ccaa-8b36-098e1f275ee1@???>:

> On 9/14/22 10:02, Luciano Mannucci wrote:
> > On Wed, 14 Sep 2022 12:49:19 +0200
> > Luciano Mannucci <luciano@???> wrote:
> >
> >>> vm.dirty_background_bytes=67108864
> >>> vm.dirty_bytes=268435456
> >>>
> >>> Maybe this additional information is helpful:
> >>>
> >>> https://forum.proxmox.com/threads/io-performance-tuning.15893/
> >>> https://lonesysadmin.net/2013/12/22/better-linux-disk-caching-performance-vm-dirty_ratio/
> >>>
> >>> Hope that helps,
> >> Yes, it does!
> >> Works like a charm!
> > I've been to quick...
> > Now only if the data comes from the local LAN (not drossing routers
> > or firewalls) I still get
> >

> >    kernel:[ 7336.007287] watchdog: BUG: soft lockup - CPU#0 stuck
> > for 22s! [swapper/0:0]

> >
> > if I write to the disk via dd nothing wrong happens...
> >
> > Luciano.
>
> Check which scheduler you are using, for virtual machine loads you
> might want to use "deadline", assuming your disk is sda, the first
> command checks your scheduler, the second changes to deadline.
>
> cat /sys/block/sda/queue/scheduler
>
> echo "deadline" >/sys/block/sda/queue/scheduler
>

..a possibly simpler idea: Give the vm another cpu thread to
help do the i/o work?

--
..med vennlig hilsen = with Kind Regards from Arnt Karlsen
...with a number of polar bear hunters in his ancestry...
Scenarios always come in sets of three:
best case, worst case, and just in case.