On Mon, 2020-05-25 at 11:13 -0400, Akshay Kumar wrote:
CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you can confirm the sender and know the content is safe.
Hi David,
io_timeout for nvme is in that gist. It's a uint 7.5 onwards.
Awesome! I missed it.
Not clear on the second part. Are you saying there are patches in RHEL AMIs that don't exist RHEL 8.2 and therefore aren't in CentOS 8.2? Or are you just pointing out that since there are no CentOS 8.2 AMIs there is some aarch64 support lagging. That's an important distinction.
Pointing out that there is a lag in the full support for some of the aarch64 instances coming out currently. IT's going to be a little while before there is full support for them, but the A1 is fully supported with the latest aarch64 images.
On Mon, May 25, 2020 at 10:08 AM David Duncan via CentOS-devel < centos-devel@centos.org> wrote:
Hi Akshay
Akshay Kumar writes:
I moved to image builder a while go to get out of the image
publishing
delays so this doesn't affect me at all. Fabian mentioned not
being
able to get a hold of anyone in marketplace. I can help with that
so
if that's a sticking point why not let someone else take a crack
at
it.
We have a team at Amazon who is actively engaged with KB, Leigh, Fabian and the gang. There is no barrier to publication that comes from the AWS Marketplace. There is a sponsored account for Marketplace and AMI delivery to which at least one of the members of the CentOS team can publish at any time they find appropriate.
There were some delays in the processes for building for the latest arm64 requiring specific patches be brought in upstream from the CentOS project. Those are mostly addressed in RHEL now, but in building your own images, you can look for the specific changes in support of those:
https://github.com/aws/aws-graviton-gettting-started/ points out some of the new issues that have been addressed and where we are looking for significant updates in the libraries. The Graviton CPU supports Arm V8.0 and includes support for CRC and crypto extensions.
And then kernel requires a few changes for full support (those are commited to RHEL and should show up in Stream between the 8.2 and 8.3 kernels
commit 18b915ac6b0ac5ba7ded03156860f60a9f16df2b Author: Dominik Brodowski linux@dominikbrodowski.net Date: Tue Oct 29 18:37:52 2019 +0100
efi/random: Treat EFI_RNG_PROTOCOL output as bootloaderrandomness and at least parts of the following for add_bootloader_randomness():
commit 428826f5358c922dc378830a1717b682c0823160 Author: Hsin-Yi Wang hsinyi@chromium.org Date: Fri Aug 23 14:24:51 2019 +0800
fdt: add support for rng-seedand set RANDOM_TRUST_BOOTLOADER in the def config.
- And don't forget to include the extended NVME timeout for the instances themselves. This is kernel dependent, so the longer, unsigned int is not available until RHEL 8. The values are
reviewed here. I looked over your gist and wondered if you were setting this in the composer blueprint.
On Mon, May 25, 2020 at 7:32 AM Karanbir Singh <
kbsingh@centos.orgmailto:kbsingh@centos.org> wrote:
On 25/05/2020 11:14, Akshay Kumar wrote:
Fabian,
If you could share some details I'd be happy to take a crack at
getting
someone at AWS Marketplace to take a look.
If you have an AMI there and can import + run, what is the major
hurdle
to using them ?
ie, why do you need the image in the marketplace from a user
standpoint ?
Regards
CentOS-devel mailing list CentOS-devel@centos.org https://lists.centos.org/mailman/listinfo/centos-devel
-- David Duncan | Working from Austin, TX Partner Solutions Architect. | +1-512-507-9268, +1-423-771- 9529 TZ=America/Chicago | He/Him A: The lost context. Q: What makes top-posted replies harder to read than bottom-posted? _______________________________________________ CentOS-devel mailing list CentOS-devel@centos.org https://lists.centos.org/mailman/listinfo/centos-devel