lp:~dannf/+git/maas-preseeds

Author: dann frazier
Author Date: 2024-03-12 13:46:50 UTC

late.sh: Re-enable MOFED on DGX H100 / hidon

jammy certification is now complete.

Signed-off-by: dann frazier <dann.frazier@canonical.com>

Author: dann frazier
Author Date: 2024-02-26 21:32:04 UTC

late.sh: Silence shellcheck warnings about "local" usage

Author: dann frazier
Author Date: 2024-02-26 21:32:04 UTC

late.sh: Silence shellcheck warnings about "local" usage

Author: dann frazier
Author Date: 2024-02-20 17:02:37 UTC

Disable MOFED installation on hidon again

So cert team can run the iperf cert test.

Author: dann frazier
Author Date: 2024-02-20 17:02:37 UTC

Disable MOFED installation on hidon again

So cert team can run the iperf cert test.

Author: dann frazier
Author Date: 2024-02-14 18:22:59 UTC

install_gpu_drivers(): Add additional kernel metapackage patterns

We're currently missing the pattern for linux-nvidia HWE kernels, which
results in us installing the DKMS package instead of the signed modules.

Author: dann frazier
Author Date: 2024-02-13 18:17:45 UTC

hidon: don't write to /home/ubuntu, it doesn't exist yet

Seen on the console:
[ 452.036488] cloud-init[4510]: + dmidecode -s system-product-name
[ 452.052372] cloud-init[4510]: + [ DGXH100 = DGXH100 ]
[ 452.068371] cloud-init[4510]: + touch /home/ubuntu/JEFF-THIS-HAS-MOFED-INSTALLED-DONT-IPERF
[ 452.088364] cloud-init[4510]: touch: cannot touch '/home/ubuntu/JEFF-THIS-HAS-MOFED-INSTALLED-DONT-IPERF': No such file or directory

Author: dann frazier
Author Date: 2024-02-13 14:50:45 UTC

hidon: Re-enable MOFED installation and try to warn the cert team

Author: dann frazier
Author Date: 2024-02-01 16:50:40 UTC

Use dmidecode to recognize hidon

$HOSTNAME apparently isn't set during deployment

Author: dann frazier
Author Date: 2024-02-01 15:27:04 UTC

Don't install MOFED on hidon until we're done with certification

I've left a reminder to undo this later here:
  https://warthogs.atlassian.net/browse/NVDGX-615

Author: dann frazier
Author Date: 2024-01-30 14:59:50 UTC

Start installing MOFED on DGX systems running 22.04 ('jammy')

Using 23.10-1.1.9.0 per:
  https://docs.google.com/spreadsheets/d/1gsRi-yTwFwdCRV2Gi_Kixsic7ZxwNdlG4P0qkyGyYfs/edit#gid=1099522443

https://warthogs.atlassian.net/browse/NVDGX-543

Author: dann frazier
Author Date: 2024-01-03 00:38:28 UTC

Add a comment to explain the use of a regex pattern for the kernel package

Author: dann frazier
Author Date: 2024-01-02 15:49:34 UTC

Add a preseed for grace-based systems

We are currently using our Grace-based systems as proxies for Grace-based
systems that we do not have in-house. Those systems always use the nvidia-
optimized kernel, so we should do the same on our systems.

This preseed will install the optimized nvidia kernel counterpart of the
generic kernel MAAS has been asked to install (since MAAS does not yet
have the ability to request an optimized kernel itself).
Install the nvidia-64k kernel by default, as this is the one we are testing

Author: dann frazier
Author Date: 2023-08-09 22:46:46 UTC

late.sh: Clear the SEL

https://warthogs.atlassian.net/browse/NVDGX-362

Author: dann frazier
Author Date: 2023-05-30 19:57:01 UTC

Use the R525 NVIDIA GPU driver branch for jammy

Ref: https://canonical.my.salesforce.com/5004K00000TIGSiQAP

Author: dann frazier
Author Date: 2023-05-30 19:57:01 UTC

Use the R525 NVIDIA GPU driver branch for jammy

Ref: https://canonical.my.salesforce.com/5004K00000TIGSiQAP

Author: dann frazier
Author Date: 2023-03-15 15:36:54 UTC

Revert "Use curtin to install linux-nvidia"

Deployments are failing with:
  E: Unable to locate package linux-nvidia

This reverts commit 1b8e33b87ecd8040182741dc82b27f28e4c067d5.

Author: dann frazier
Author Date: 2023-02-14 18:58:04 UTC

Add symlinks for cortez

Author: dann frazier
Author Date: 2022-03-30 15:07:01 UTC

curtin_userdata_dgx: Disable dynamic storage configuration

Currently broken in MAAS 3.2 beta1:
  https://bugs.launchpad.net/maas/+bug/1967008

Author: dann frazier
Author Date: 2021-12-13 21:53:47 UTC

Disable knem autoloading to avoid oops (LP: #1929187)

Author: dann frazier
Author Date: 2021-12-09 14:43:08 UTC

late.sh: Require the A100 to be in NPS=4 mode now

We're now running performance benchmarks in NPS=4 mode, which is how
these systems are configured from the factory.

Author: dann frazier
Author Date: 2021-11-04 01:05:16 UTC

m400: drop use of workaround PPA for LP: #1926985

The equivalent SRUs have now been released into the archive. LP: #1949048

Author: dann frazier
Author Date: 2021-10-22 20:55:21 UTC

Workaround issue w/ knem-dkms

Upon install it only builds against the running kernel, which is not
necessarily the installed kernel.

http://partners.nvidia.com/bug/viewbug/3409217