Date: Sun, 11 Feb 2024 20:10:05 -0800 From: "Chuck Tuffli" <chuck@tuffli.net> To: "Jason Tubnor" <jason@tubnor.net>, "FreeBSD virtualization" <freebsd-virtualization@freebsd.org>, "Matthew Grooms" <mgrooms@shrew.net> Subject: Re: 14R bhyve RHEL9+NVME Message-ID: <6c9bd61c-a3ca-4470-895c-494297af0956@app.fastmail.com> In-Reply-To: <20240212012559.Horde.aH5FMCkNv4j7GfULz1fq5rn@drive.tubnor.net> References: <678ac1d2-d00c-450e-bc2e-f9d979d19340@shrew.net> <20240212012559.Horde.aH5FMCkNv4j7GfULz1fq5rn@drive.tubnor.net>
next in thread | previous in thread | raw e-mail | index | archive | help
--d7044c208f3a48baa3daca6471f24497
Content-Type: text/plain;charset=utf-8
Content-Transfer-Encoding: quoted-printable
On Sun, Feb 11, 2024, at 5:25 PM, Jason Tubnor wrote:
> Hi Matthew,
> =20
> There is some underlying bug post 9.1 when it comes to Anaconda and th=
e presented NVMe storage. I can't reproduce it with newer version of Ana=
conda that is present in Fedora.
> =20
> Uplifting in place from 9.1 to 9.3, there are no issues once the opera=
ting system is up and running, so it isn't a RHEL thing, just the instal=
ler.
> =20
> Cheers,
> =20
> Jason.
> =20
> "Matthew Grooms" mgrooms@shrew.net =E2=80=93 12 February 2024 11:32 AM
> > Hey All,
> >=20
> > Has anyone had experience installing either Rocky or Alma Linux 9.3=20
> > using virtual NVME disks on a FreeBSD 14R host? I have no problem us=
ing=20
> > VirtIO block devices or if I downgrade to an 8.9 variant. However, u=
sing=20
> > NVME with a 9.3 variant will cause the installer to consistently fai=
l on=20
> > my system. I'm not sure how to gather more data at this point as the=20
> > installer displays a wall of text that I'm not sure how to extract. =
I've=20
> > already burned a few hours on this so I thought I'd reach out to the=20
> > list to see if anyone else has already dug into this. If not, I'll s=
ee=20
> > what I can do to gather more info. For reference, here are the ISOs =
I've=20
> > tried installing from so far ...
I have heard that anaconda has a debug mode activated by some kernel boo=
t option. Maybe that could shed some light.=20
>From the NVMe side, there aren=E2=80=99t any errors or =E2=80=9Cweird=E2=
=80=9D commands. My guess is the emulation is returning something anacon=
da doesn=E2=80=99t like, but it=E2=80=99s hard to tell from the Python e=
xception what that might be.=20
Note the text/rescue installer uses tmux which gives better visibility i=
nto the logs.=20
--d7044c208f3a48baa3daca6471f24497
Content-Type: text/html;charset=utf-8
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE html><html><head><title></title><style type=3D"text/css">p.Mso=
Normal,p.MsoNoSpacing{margin:0}</style></head><body><div><br></div><div>=
<br></div><div>On Sun, Feb 11, 2024, at 5:25 PM, Jason Tubnor wrote:<br>=
</div><blockquote type=3D"cite" id=3D"qt" style=3D""><div>Hi Matthew,<br=
></div><div> <br></div><div>There is some underlying bug post 9.1 w=
hen it comes to Anaconda and the presented NVMe storage. I can't reprodu=
ce it with newer version of Anaconda that is present in Fedora.<br></div=
><div> <br></div><div>Uplifting in place from 9.1 to 9.3, there are=
no issues once the operating system is up and running, so it isn't a RH=
EL thing, just the installer.<br></div><div> <br></div><div>Cheers,=
<br></div><div> <br></div><div>Jason.<br></div><div> <br></div=
><div>"Matthew Grooms" <a href=3D"mailto:mgrooms@shrew.net" class=3D=
"">mgrooms@shrew.net</a> =E2=80=93 12 February 2024 11:32 AM<br></div><d=
iv>> Hey All,<br></div><div>> <br></div><div>> Has anyone =
had experience installing either Rocky or Alma Linux 9.3 <br></div>=
<div>> using virtual NVME disks on a FreeBSD 14R host? I have no prob=
lem using <br></div><div>> VirtIO block devices or if I downgrad=
e to an 8.9 variant. However, using <br></div><div>> NVME with a=
9.3 variant will cause the installer to consistently fail on <br><=
/div><div>> my system. I'm not sure how to gather more data at this p=
oint as the <br></div><div>> installer displays a wall of text t=
hat I'm not sure how to extract. I've <br></div><div>> already b=
urned a few hours on this so I thought I'd reach out to the <br></d=
iv><div>> list to see if anyone else has already dug into this. If no=
t, I'll see <br></div><div>> what I can do to gather more info. =
For reference, here are the ISOs I've <br></div><div>> tried ins=
talling from so far ...<br></div></blockquote><div><br></div><div>I have=
heard that anaconda has a debug mode activated by some kernel boot opti=
on. Maybe that could shed some light. <br></div><div><br></div><div>From=
the NVMe side, there aren=E2=80=99t any errors or =E2=80=9Cweird=E2=80=9D=
commands. My guess is the emulation is returning something anaconda doe=
sn=E2=80=99t like, but it=E2=80=99s hard to tell from the Python excepti=
on what that might be. <br></div><div><br></div><div>Note the text/=
rescue installer uses tmux which gives better visibility into the logs.&=
nbsp;</div></body></html>
--d7044c208f3a48baa3daca6471f24497--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?6c9bd61c-a3ca-4470-895c-494297af0956>