Your message dated Sat, 12 Oct 2024 19:00:19 +0200
with message-id <zwqri_dpsuf3s...@eldamar.lan>
and subject line Re: Bug#1082081: Problem actually unreproducible for us
has caused the Debian Bug report #1082081,
regarding Performance issues on VM (virtio)
to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact ow...@bugs.debian.org
immediately.)
--
1082081: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1082081
Debian Bug Tracking System
Contact ow...@bugs.debian.org with problems
--- Begin Message ---
Package: linux-image-6.1.0-24-cloud-amd64
Version: 6.1.106-2
Severity: serious
Tags: upstream
Forwarded: https://bugzilla.kernel.org/show_bug.cgi?id=219129
After the latest linux upgrade, the following message appear frequently
in /var/log/syslog:
ens3: bad gso: type: 4, size: 1460
It seems real network issues have been noticed in some VMs we rebooted
on friday (they are now running the previous kernel).
The forwarded upstream issue seems to match.
Regards
taffit
signature.asc
Description: PGP signature
--- End Message ---
--- Begin Message ---
Source: linux
Source-Version: 6.1.112-1
On Tue, Oct 08, 2024 at 08:00:27AM +0000, Kocher Emanuel, Bedag wrote:
> Hi Salvatore,
>
> On 04/10/2024 22:24, Salvatore Bonaccorso wrote:
> > Hi Emanuel,
> >
> > On Wed, Sep 25, 2024 at 08:59:00PM +0000, Kocher Emanuel, Bedag wrote:
> >> Hi David,
> >>
> >> On 25/09/2024 15:35, David Prevot wrote:
> >>
> >> Hi,
> >>
> >> After further investigations, I now believe we’re not able to reproduce
> >> the issue (the performance issues were actually random, and have other
> >> causes that we’re currently investigating).
> >>
> >> Emanuel, have you been able to test a kernel including the patch
> >> previously pointed by Salvatore and Florian?
> >>
> >> 34aaedb052cb4cb068cb19a34cf6d9cedb9af87f
> >> https://lore.kernel.org/all/20240910213553.839926-1-willemdebruijn.ker...@gmail.com/
> >>
> >> I’ve built it last week following the team documentation.
> >>
> >> https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#id-1.6.6.4
> >>
> >> It’s available at the following URL if needed.
> >>
> >> https://download.evolix.net/b3e804f0aaf78b4dc217213942c20805079fd160/linux-image-6.1.0-0.a.test-amd64-unsigned_6.1.106-3a~test_amd64.deb
> >>
> >>
> >> This is unfortunate to hear - we actually still have the issue with the
> >> patched kernel and can't see any improvements at all with speed nor
> >> reliability.
> >>
> >> Even a simple copy of the linux kernel from a nfs share to the homedir
> >> will interrupt after a while with an "Input/output error".
> >>
> >> Turning GSO off did not help as some suggested:
> >> echo options virtio_net gso=N >> /etc/modprobe.d/virtio.conf
> >> rmmod virtio_net && modprobe virtio_net
> >>
> >> As for turning LSO off on the hypervisor I would have to get a test host -
> >> it seems to have helped here:
> >> https://github.com/virtio-win/kvm-guest-drivers-windows/issues/1131
> >>
> >> If there's something we can do in order of testing we should be able to do
> >> so.
> > Since you still are affected by the issue: Can you please the newes
> > 6.1.112-1 which was released as DSA 5782-1?
>
> I can confirm that 6.1.112-1 resolves the issue and the network
> performance is back to normal.
>
> Tested on different virtual machines and upgrading the rest now, thank
> you and the team!
Many thanks for the confirmation! Closing the bug with the version
now.
Regards,
Salvatore
--- End Message ---
