[cfarm-users] Future of cfarm186/cfarm187/cfarm188 at OSUOSL

[Collin Funk]

Jeffrey Walton via cfarm-users <cfarm-users at lists.tetaneutral.net>
writes:

> On Sat, Nov 15, 2025 at 7:10 AM Baptiste Jonglez via cfarm-users
> <cfarm-users at lists.tetaneutral.net> wrote:
>>
>> As some of you may know, OSUOSL is moving to a new datacenter.
>>
>> All "exotic" hosts (POWER etc) will be moved, it should be transparent
>> except for a bit of downtime.
>>
>> However, the three x86 machines (cfarm186, 187, 188) are getting quite
>> old, so OSUOSL is looking at replacing them with newer hardware.  We
>> probably won't be able to keep as many machines.
>>
>> Here are two possible options to start the discussion:
>>
>> - option 1: one much bigger machine, bare metal.  Which (single) OS makes the most sense?
>>
>> - option 2: several virtual machines.  Which (multiple) OS would be useful?
>
> I think Hurd would make a good choice to help with portability of
> C/C++ programs for a couple of reasons.
>
> First, Hurd lacks PATH_MAX and friends, and it is a surprise to some
> program authors, and a pain in the butt for package maintainers and
> porters.  Also see
> <https://www.gnu.org/software/hurd/community/gsoc/project_ideas/maxpath.html>.
>
> Second, setting up a Hurd virtual machine has some sharp edges.  It is
> not as easy as setting up a Debian or Red Hat vm.  If a vm was already
> available, it could help folks interested in porting to Hurd since the
> machine is already available.

I'm not sure how feasible it would be to run Hurd. I have crashed it a
few times while testing Coreutils, for example [1].

It is pretty simple to setup Qemu VMs though, luckily. The ISOs can be
found here [2]. And instructions can be found in the README [3].

Collin

[1] https://github.com/coreutils/coreutils/commit/8dd89f55d52059e590572ed94a8c9165f92b5354
[2] https://cdimage.debian.org/cdimage/ports/13.0/hurd-amd64/
[3] https://cdimage.debian.org/cdimage/ports/13.0/hurd-amd64/README.txt