[LEP-3777] fix(nvidia): do not exit on NVML init error (instead returns Unhealthy)#1180
Merged
Conversation
gyuho
force-pushed
the
handle-nvml-error
branch
from
December 29, 2025 05:24
0dc9cde to
2aa2bfe
Compare
gyuho
force-pushed
the
handle-nvml-error
branch
from
January 8, 2026 04:54
16deecb to
f175aa8
Compare
eahydra
reviewed
Jan 8, 2026
…ns Unhealthy) Signed-off-by: Gyuho Lee <gyuhol@nvidia.com>
gyuho
force-pushed
the
handle-nvml-error
branch
from
January 12, 2026 10:36
f175aa8 to
189a250
Compare
…nents Extends the graceful NVML initialization error handling from clock-speed (commit 56024b2) to all 18 remaining nvidia accelerator components. Changes: - Add InitError() check to all component Check() methods - Add InitError() method to all test mock NVML instances - Add suggestedActions field to 5 component checkResult structs (fabric-manager, nccl, peermem, sxid, xid) - Add missing fmt import to infiniband component When NVML initialization fails (e.g., "error getting device handle for index 'N': Unknown Error" / "GPU has fallen off the bus"), components now report unhealthy status with reboot suggestion instead of causing gpud to exit. This ensures gpud remains operational while clearly reporting GPU hardware issues that require system intervention. Signed-off-by: Gyuho Lee <gyuhol@nvidia.com>
gyuho
force-pushed
the
handle-nvml-error
branch
from
January 12, 2026 16:51
189a250 to
dd08a68
Compare
eahydra
approved these changes
Jan 13, 2026
Signed-off-by: Gyuho Lee <gyuhol@nvidia.com>
gyuho
added a commit
that referenced
this pull request
Jan 15, 2026
… testing NVML init failures This adds a hidden testing flag to simulate Device().GetDevices() failures, enabling integration testing of the erroredInstance path introduced in: #1180 The flag simulates the "Unable to determine the device handle for GPU: Unknown Error" scenario that occurs when NVML loads but device enumeration fails (e.g., Xid 79). When enabled, gpud continues running but all nvidia components report unhealthy. Signed-off-by: Gyuho Lee <gyuhol@nvidia.com>
gyuho
added a commit
that referenced
this pull request
Jan 15, 2026
… testing NVML init failures This adds a hidden testing flag to simulate Device().GetDevices() failures, enabling integration testing of the erroredInstance path introduced in: #1180 The flag simulates the "Unable to determine the device handle for GPU: Unknown Error" scenario that occurs when NVML loads but device enumeration fails (e.g., Xid 79). When enabled, gpud continues running but all nvidia components report unhealthy. Signed-off-by: Gyuho Lee <gyuhol@nvidia.com>
gyuho
added a commit
that referenced
this pull request
Jan 16, 2026
… testing NVML init failures This adds a hidden testing flag to simulate Device().GetDevices() failures, enabling integration testing of the erroredInstance path introduced in: #1180 The flag simulates the "Unable to determine the device handle for GPU: Unknown Error" scenario that occurs when NVML loads but device enumeration fails (e.g., Xid 79). When enabled, gpud continues running but all nvidia components report unhealthy. Signed-off-by: Gyuho Lee <gyuhol@nvidia.com>
gyuho
added a commit
that referenced
this pull request
Jan 19, 2026
… testing NVML init failures (#1196) This adds a hidden testing flag to simulate Device().GetDevices() failures, enabling integration testing of the erroredInstance path introduced in: #1180 The flag simulates the "Unable to determine the device handle for GPU: Unknown Error" scenario that occurs when NVML loads but device enumeration fails (e.g., Xid 79). When enabled, gpud continues running but all nvidia components report unhealthy. Signed-off-by: Gyuho Lee <gyuhol@nvidia.com> Signed-off-by: Gyuho Lee <gyuhol@nvidia.com>
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
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.
Signed-off-by: Gyuho Lee gyuhol@nvidia.com