Add draft gpu troubles #290
Draft
Conversation
This file contains 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
perifaws
suggested changes
Apr 30, 2024
| scancel [JOB_ID] | ||
| ``` | ||
|
|
||
| 1. Reset the GPUs |
There was a problem hiding this comment.
Add a link to the reset option for nvidia-smi
| sudo /opt/slurm/bin/scontrol create res starttime=now duration=infinite flags=ignore_jobs user=root nodes=[NODE_TO_TERMINATE] | ||
| ``` | ||
|
|
||
| 1. Cancel |
| The node will have a **DRAIN** status. Then the instance will be terminated and replaced. | ||
|
|
||
|
|
||
| 1. Delete the reservation |
| scancel [JOB_ID] | ||
| ``` | ||
|
|
||
| 1. Place the node in **DRAIN**. |
There was a problem hiding this comment.
node to terminate is an IP or name?
|
|
||
| 1. Create a reservation to isolate the node from being used by any jobs. | ||
| ```bash | ||
| sudo /opt/slurm/bin/scontrol create res starttime=now duration=infinite flags=ignore_jobs user=root nodes=[NODE_TO_TERMINATE] |
There was a problem hiding this comment.
what NODE_TO_TERMINATE should be?
| | 95 | Uncontained ECC error | Reset GPUs | [AWS ParallelCluster](#reset-gpus) | | ||
|
|
||
| # AWS ParallelCluster | ||
|
|
There was a problem hiding this comment.
reference to ParallelCluster doc?
| While running High-Performance Computing or Machine Learning workloads, GPUs may fail for various reasons captured by Xid messages. | ||
| Those messages are placed in `/var/log/messages` for Amazon Linux or for Ubuntu in `/var/log/syslog` and `/var/log/kern.log` | ||
|
|
||
| | Xid | Failure | Resolution | Orchestrator | |
| ``` | ||
|
|
||
| ## Reset GPUs | ||
|
|
There was a problem hiding this comment.
Say that resetting does and what NODE_TO_TERMINATE represents (or how to get it)
|
|
||
| 1. Delete the reservation | ||
| ```bash | ||
| sudo /opt/slurm/bin/scontrol delete res root_[RES_NUMBER] |
|
|
||
| # Amazon SageMaker HyperPod | ||
|
|
||
| TBD |
|
looks good - plan to create a new PR for HyperPod instructions after p-cluster is merged. |
KeitaW
force-pushed
the
feature/#289_gpu_failure
branch
2 times, most recently
from
c9e0f5e to
30e6592
Compare
KeitaW
force-pushed
the
main
branch
3 times, most recently
from
44e448e to
1209815
Compare
mhuguesaws
force-pushed
the
feature/#289_gpu_failure
branch
2 times, most recently
from
90549b2 to
84f6f79
Compare
Signed-off-by: AWS ParallelCluster user <[email protected]>
Signed-off-by: AWS ParallelCluster user <[email protected]>
mhuguesaws
force-pushed
the
feature/#289_gpu_failure
branch
from
84f6f79 to
7a102a3
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Issue #, if available:
Description of changes:
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.