Initial fix for non-detection of OmniPath cards #387
Merged
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
…eus#341) Currently these cards aren't picked up correctly because the driver doesn't provide the hca_type file the parse function expects. This patch stops the parse function treating that as an error. Signed-off-by: Ian Kirker <[email protected]>
This adds fixtures and a test for the OmniPath detection and metric reading added in a previous commit. This could use a check from someone better acquainted with the driver and relevant metrics, but it should do some approximation of the job for now. It might even be sensible to specifically detect major versions of devices, since the detailed metric file layout might even vary between mlx4 and mlx5. There's also a factor of 4 included in the Mellanox metric pickup that I'm not sure about: might need to detect device to tell whether we need that, and that's a bit more work. Signed-off-by: Ian Kirker <[email protected]>
|
Ah. I had to do some finicky rebasing + merging, and I think that's where the test failures come from -- I tested before but not after, foolishly. I'll fix it up. |
Whoops. Signed-off-by: Ian Kirker <[email protected]>
remijouannet
pushed a commit
to remijouannet/procfs
that referenced
this pull request
Oct 20, 2022
* Fixes InfiniBand fileset incompatibility with OmniPath cards (prometheus#341) Currently these cards aren't picked up correctly because the driver doesn't provide the hca_type file the parse function expects. This patch stops the parse function treating that as an error. Signed-off-by: Ian Kirker <[email protected]> * Adds test for OmniPath device detection This adds fixtures and a test for the OmniPath detection and metric reading added in a previous commit. This could use a check from someone better acquainted with the driver and relevant metrics, but it should do some approximation of the job for now. It might even be sensible to specifically detect major versions of devices, since the detailed metric file layout might even vary between mlx4 and mlx5. There's also a factor of 4 included in the Mellanox metric pickup that I'm not sure about: might need to detect device to tell whether we need that, and that's a bit more work. Signed-off-by: Ian Kirker <[email protected]> * Fixes some mistakes that slipped in during merge/rebase Whoops. Signed-off-by: Ian Kirker <[email protected]>
jritter
pushed a commit
to jritter/procfs
that referenced
this pull request
Jul 15, 2024
* Fixes InfiniBand fileset incompatibility with OmniPath cards (prometheus#341) Currently these cards aren't picked up correctly because the driver doesn't provide the hca_type file the parse function expects. This patch stops the parse function treating that as an error. Signed-off-by: Ian Kirker <[email protected]> * Adds test for OmniPath device detection This adds fixtures and a test for the OmniPath detection and metric reading added in a previous commit. This could use a check from someone better acquainted with the driver and relevant metrics, but it should do some approximation of the job for now. It might even be sensible to specifically detect major versions of devices, since the detailed metric file layout might even vary between mlx4 and mlx5. There's also a factor of 4 included in the Mellanox metric pickup that I'm not sure about: might need to detect device to tell whether we need that, and that's a bit more work. Signed-off-by: Ian Kirker <[email protected]> * Fixes some mistakes that slipped in during merge/rebase Whoops. Signed-off-by: Ian Kirker <[email protected]>
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.
Mellanox drivers provide a particular file in the
/syspseudo-filesystem that not all InfiniBand drivers do, and OmniPath drivers don't. This means that the InfiniBand detection fails to detect these cards.This naïve fix makes OmniPath cards detectable by making the InfiniBand detection happy to ignore whether that file is missing.
It could probably use a look in future from someone with more expertise (or at least, more documentation) on the specific info provided by the driver.
This fixes #341 and I think also the downstream issue in the node_exporter: prometheus/node_exporter#2023.