Skip to content

Update Azure GRID driver to 580.178.04 - #4332

Merged
peterschmidt85 merged 2 commits into
masterfrom
pr_azure_grid_driver_580
Oct 2, 2026
Merged

peterschmidt85 merged 2 commits into
masterfrom
pr_azure_grid_driver_580

Conversation

@peterschmidt85

@peterschmidt85 peterschmidt85 commented Oct 2, 2026 •

Copy link
Copy Markdown
Contributor

NVIDIA's 2026-09-30 security bulletin covers A10 vGPU, and Azure now supports only the patched GRID guest drivers on NVadsA10_v5 VMs. The dstack-grid image shipped 570.211.01, which is no longer supported and lacks CUDA 13 support, so CUDA 13 images without forward compatibility libraries failed on A10 with "driver too old".

  • Update the GRID driver in the Azure dstack-grid image to 580.178.04 (vGPU 19.6 LTS), which supports CUDA 13.0 like the dstack-cuda images.
  • Add azure_variant to the VM images workflow to build only some Azure image variants.
  • Add azure_new_gallery_version to republish an existing Azure image as the next gallery version. dstack uses the latest gallery version, so servers on dstack-grid-0.14 get the new driver without upgrading. Without it, publishing stops if the image already has a version.

🤖 Generated with Claude Code

Andrey Cheptsov and others added 2 commits October 2, 2026 12:47
NVIDIA's 2026-09-30 security bulletin covers A10 vGPU. Azure has patched
the NVadsA10_v5 hosts and now supports only the patched guest drivers:
vGPU 19.6 (580.178.04) and vGPU 20.2.

The 580 driver also adds CUDA 13.0 support on A10. With the previous 570
driver, CUDA 13 images failed with "driver too old" unless they bundled
CUDA forward compatibility libraries, as dstack's base image does. This
matches the 580 driver already used in dstack-cuda images.

Also allow republishing an existing Azure image as a new gallery version,
so dstack servers using dstack-grid-0.14 pick up the new driver without
a release:
- azure_variant input to build only some Azure image variants
- azure_new_gallery_version input to replace the build image and publish
  the next gallery version; without it, publishing stops if a version exists

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Without this, a failed version listing looked like an empty gallery, so
the script tried to publish 0.0.1 again instead of the next version.
Also describe azure_new_gallery_version as a rebuild, since packer
-force deletes and rebuilds the existing managed image.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@peterschmidt85
peterschmidt85 merged commit 90887d0 into master Oct 2, 2026
30 checks passed
@peterschmidt85
peterschmidt85 deleted the pr_azure_grid_driver_580 branch October 2, 2026 13:47
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant