With the current efforts under way with aarch64 and ppc64le, I would like to suggest that s390x also be added as an alternative architecture. Obviously hardware would be the biggest obstacle, but I do not see any reason why we couldn't get started with just emulated hardware.
On Jul 10 10:43, Jason DeTiberus wrote:
With the current efforts under way with aarch64 and ppc64le, I would like to suggest that s390x also be added as an alternative architecture. Obviously hardware would be the biggest obstacle, but I do not see any reason why we couldn't get started with just emulated hardware.
-- Jason DeTiberus
CentOS-devel mailing list CentOS-devel@centos.org https://lists.centos.org/mailman/listinfo/centos-devel
What sort of hardware do we need to support the emulation? I'd imagine this would be painfully slow for the bootstrap, but perhaps that's ok.
Is there a group of folks interested enough in s390x to give pointers and/or reviews on the admin side of things?
-- Brian Stinson
On Mon, Jul 10, 2017 at 10:54 AM, Brian Stinson brian@bstinson.com wrote:
On Jul 10 10:43, Jason DeTiberus wrote:
With the current efforts under way with aarch64 and ppc64le, I would like to suggest that s390x also be added as an alternative architecture. Obviously hardware would be the biggest obstacle, but I do not see any reason why we couldn't get started with just emulated hardware.
-- Jason DeTiberus
CentOS-devel mailing list CentOS-devel@centos.org https://lists.centos.org/mailman/listinfo/centos-devel
What sort of hardware do we need to support the emulation? I'd imagine this would be painfully slow for the bootstrap, but perhaps that's ok.
libvirt/qemu supports s390x emulation, so we could simply use commodity x86_64 hardware.
On 10/07/17 15:43, Jason DeTiberus wrote:
With the current efforts under way with aarch64 and ppc64le, I would like to suggest that s390x also be added as an alternative architecture. Obviously hardware would be the biggest obstacle, but I do not see any reason why we couldn't get started with just emulated hardware.
First thing we would need to get going is the distro bootstrap - given that 7.x is now almost 3 years in - we will need to go back to the pre release beta and bootstrap up from there.
That also means we will need the corresponding fedora bits - which should be available.
Once we have that in place, we'll need to workout how best to get this supported in cbs.centos.org, before moving to ci.centos.org support.
I'd expected this to be quite a major commitment from someone over a couple of months, given that hardware to back this in emulated mode wont be a huge problem - it just wont be quick.
On Jul 13, 2017, at 6:08 AM, Karanbir Singh kbsingh@centos.org wrote:
On 10/07/17 15:43, Jason DeTiberus wrote:
With the current efforts under way with aarch64 and ppc64le, I would like to suggest that s390x also be added as an alternative architecture. Obviously hardware would be the biggest obstacle, but I do not see any reason why we couldn't get started with just emulated hardware.
First thing we would need to get going is the distro bootstrap - given that 7.x is now almost 3 years in - we will need to go back to the pre release beta and bootstrap up from there.
That also means we will need the corresponding fedora bits - which should be available.
Once we have that in place, we'll need to workout how best to get this supported in cbs.centos.org, before moving to ci.centos.org support.
I'd expected this to be quite a major commitment from someone over a couple of months, given that hardware to back this in emulated mode wont be a huge problem - it just wont be quick.
I would be willing to throw my hat into ring to help with the s390x port. Most of my ppc64le/ppc64 porting memory cells are still intact.
Having real s390x builder vms like fedora would help https://koji.fedoraproject.org/koji/hosts?start=100&state=enabled&or...
-James
On Jul 13 12:59, James O'Connor wrote:
On Jul 13, 2017, at 6:08 AM, Karanbir Singh kbsingh@centos.org wrote:
On 10/07/17 15:43, Jason DeTiberus wrote:
With the current efforts under way with aarch64 and ppc64le, I would like to suggest that s390x also be added as an alternative architecture. Obviously hardware would be the biggest obstacle, but I do not see any reason why we couldn't get started with just emulated hardware.
First thing we would need to get going is the distro bootstrap - given that 7.x is now almost 3 years in - we will need to go back to the pre release beta and bootstrap up from there.
That also means we will need the corresponding fedora bits - which should be available.
Once we have that in place, we'll need to workout how best to get this supported in cbs.centos.org, before moving to ci.centos.org support.
I'd expected this to be quite a major commitment from someone over a couple of months, given that hardware to back this in emulated mode wont be a huge problem - it just wont be quick.
I would be willing to throw my hat into ring to help with the s390x port. Most of my ppc64le/ppc64 porting memory cells are still intact.
Having real s390x builder vms like fedora would help https://koji.fedoraproject.org/koji/hosts?start=100&state=enabled&or...
-James _______________________________________________ CentOS-devel mailing list CentOS-devel@centos.org https://lists.centos.org/mailman/listinfo/centos-devel
I think we're pretty squarely in the emulation side of things. Access to 'real' s390x is hard to come by.
If there is a group committed to the bootstrap, let's get together and talk about getting some resources assigned to this. Anyone else want in?
--Brian
On Thu, Jul 13, 2017 at 2:06 PM, Brian Stinson brian@bstinson.com wrote:
On Jul 13 12:59, James O'Connor wrote:
On Jul 13, 2017, at 6:08 AM, Karanbir Singh kbsingh@centos.org
wrote:
On 10/07/17 15:43, Jason DeTiberus wrote:
With the current efforts under way with aarch64 and ppc64le, I would like to suggest that s390x also be added as an alternative
architecture.
Obviously hardware would be the biggest obstacle, but I do not see any reason why we couldn't get started with just emulated hardware.
First thing we would need to get going is the distro bootstrap - given that 7.x is now almost 3 years in - we will need to go back to the pre release beta and bootstrap up from there.
That also means we will need the corresponding fedora bits - which should be available.
Once we have that in place, we'll need to workout how best to get this supported in cbs.centos.org, before moving to ci.centos.org support.
I'd expected this to be quite a major commitment from someone over a couple of months, given that hardware to back this in emulated mode
wont
be a huge problem - it just wont be quick.
I would be willing to throw my hat into ring to help with the s390x
port. Most of my ppc64le/ppc64 porting memory cells are still intact.
Having real s390x builder vms like fedora would help
https://koji.fedoraproject.org/koji/hosts?start=100& state=enabled&order=name
-James _______________________________________________ CentOS-devel mailing list CentOS-devel@centos.org https://lists.centos.org/mailman/listinfo/centos-devel
I think we're pretty squarely in the emulation side of things. Access to 'real' s390x is hard to come by.
I've recently become aware of the Open Mainframe Project ( https://www.openmainframeproject.org/). I'm wondering if there are possible collaboration efforts there that could help with hardware access.
If there is a group committed to the bootstrap, let's get together and talk about getting some resources assigned to this. Anyone else want in?
If it wasn't obvious from my initial request, I am interested in helping with the bootstrapping efforts as well.
On 14/07/17 17:15, Jason DeTiberus wrote:
On Thu, Jul 13, 2017 at 2:06 PM, Brian Stinson <brian@bstinson.com mailto:brian@bstinson.com> wrote:
On Jul 13 12:59, James O'Connor wrote: >
<snip>
I've recently become aware of the Open Mainframe Project (https://www.openmainframeproject.org/). I'm wondering if there are possible collaboration efforts there that could help with hardware access.
Well, as we're talking about building the distro, the policy for this (from a CentOS side) was that from a security/auditing PoV we always said that distro build should happen on machine under our control, and so hosted.
KB,
Should we spin up a formal SIG around this?
-- Jason DeTiberus
On Fri, Jul 14, 2017 at 11:21 AM, Fabian Arrotin arrfab@centos.org wrote:
On 14/07/17 17:15, Jason DeTiberus wrote:
On Thu, Jul 13, 2017 at 2:06 PM, Brian Stinson <brian@bstinson.com mailto:brian@bstinson.com> wrote:
On Jul 13 12:59, James O'Connor wrote: ><snip> > > I've recently become aware of the Open Mainframe Project > (https://www.openmainframeproject.org/). I'm wondering if there are > possible collaboration efforts there that could help with hardware access. >
Well, as we're talking about building the distro, the policy for this (from a CentOS side) was that from a security/auditing PoV we always said that distro build should happen on machine under our control, and so hosted.
-- Fabian Arrotin The CentOS Project | http://www.centos.org gpg key: 56BEC54E | twitter: @arrfab
CentOS-devel mailing list CentOS-devel@centos.org https://lists.centos.org/mailman/listinfo/centos-devel
-
Brian Stinson -
Fabian Arrotin -
James O'Connor -
Jason DeTiberus -
Karanbir Singh