AZP: Accept nvidia_peermem and drop legacy service check #10887
+0
−1
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
- Pass GPU peer-memory check if either nvidia_peermem or nv_peer_mem is loaded - Remove checking for old nv_peer_mem systemd service - Keep try_load_cuda_env() behavior: verify via /sys/kernel/mm/memory_peers/nv_mem/version - Remove redundant lsmod grep and stop referencing nv_peer_mem systemd service - Support hosts migrating to nvidia_peermem while preserving backward compatibility Signed-off-by: Alexey Rivkin <[email protected]>
AZP: accept nvidia_peermem and drop legacy service check Signed-off-by: Alexey Rivkin <[email protected]>
dpressle
requested changes
Sep 15, 2025
| if ! lsmod | grep -q 'nv.*_peer.*mem'; then | ||
| lsmod | grep 'nv.*_peer.*mem' | ||
| systemctl status nv_peer_mem | ||
| azure_log_error "nv_peer_mem module not loaded on $(hostname -s)" |
There was a problem hiding this comment.
can you change the message to say nv_peer_mem|nvidia_peermem
There was a problem hiding this comment.
Maybe just "NV peer memory module not loaded on $(hostname -s)"?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What?
nvidia_peermemalongsidenv_peer_memin GPU peer-memory checks inbuildlib/az-helpers.sh.nv_peer_memsystemd service.try_load_cuda_env()verifying via/sys/kernel/mm/memory_peers/nv_mem/version.Why?
nvidia_peermem. This avoids false negatives and removes dependency on a legacy service while preserving backward compatibility.How?
check_nv_peer_mem(): pass if eithernvidia_peermemornv_peer_memis loaded; drop service status check.try_load_cuda_env(): unchanged behavior, still relies on the sysfs version file.