Merge pull request #4418 from piyush-jena/roll-fwd

Add support for NVIDIA MIG

CHANGELOG.md

@@ -1,3 +1,20 @@
+# v1.34.0 (2025-03-03)
+
+## OS Changes
+
+* Add support for NVIDIA Multi-Instance GPU (MIG) ([#4418])
+
+## Build Changes
+* Update bottlerocket-core-kit from 6.0.1 to 6.0.2 [CHANGELOG](/~https://github.com/bottlerocket-os/bottlerocket-core-kit/blob/develop/CHANGELOG.md#v602-2025-02-26) ([#4416])
+* Update bottlerocket-kernel-kit from 1.1.2 to 1.2.0 [CHANGELOG](/~https://github.com/bottlerocket-os/bottlerocket-kernel-kit/blob/develop/CHANGELOG.md#v120-2024-02-26) ([#4416])
+
+### Twoliter
+
+* Update Twoliter to 0.7.3 ([#4416])
+
+[#4416]: /~https://github.com/bottlerocket-os/bottlerocket/pull/4416
+[#4418]: /~https://github.com/bottlerocket-os/bottlerocket/pull/4418
+
 # v1.33.0 (2025-02-26)
 
 ## Release Highlights

Release.toml

@@ -1,4 +1,4 @@
-version = "1.33.0"
+version = "1.34.0"
 
 [migrations]
 "(0.3.1, 0.3.2)" = ["migrate_v0.3.2_admin-container-v0-5-0.lz4"]
@@ -404,3 +404,6 @@ version = "1.33.0"
     "migrate_v1.33.0_public-remove-source-control.lz4",
     "migrate_v1.33.0_remove-metadata-and-weak-settings-migration.lz4",
 ]
+"(1.33.0, 1.34.0)" = [
+    "migrate_v1.34.0_kubelet-device-plugins-mig-settings.lz4",
+]

Twoliter.toml

@@ -1,5 +1,5 @@
 schema-version = 1
-release-version = "1.33.0"
+release-version = "1.34.0"
 
 [vendor.bottlerocket]
 registry = "public.ecr.aws/bottlerocket"

sources/Cargo.toml

@@ -88,6 +88,7 @@ members = [
     "settings-migrations/v1.33.0/public-remove-source-admin",
     "settings-migrations/v1.33.0/public-remove-source-control",
     "settings-migrations/v1.33.0/remove-metadata-and-weak-settings-migration",
+    "settings-migrations/v1.34.0/kubelet-device-plugins-mig-settings",
 
     "settings-plugins/aws-dev",
     "settings-plugins/aws-ecs-1",
@@ -164,22 +165,22 @@ version = "0.1.0"
 
 [workspace.dependencies.bottlerocket-modeled-types]
 git = "/~https://github.com/bottlerocket-os/bottlerocket-settings-sdk"
-tag = "bottlerocket-settings-models-v0.7.0"
-version = "0.7.0"
+tag = "bottlerocket-settings-models-v0.8.0"
+version = "0.8.0"
 
 [workspace.dependencies.bottlerocket-settings-models]
 git = "/~https://github.com/bottlerocket-os/bottlerocket-settings-sdk"
-tag = "bottlerocket-settings-models-v0.7.0"
-version = "0.7.0"
+tag = "bottlerocket-settings-models-v0.8.0"
+version = "0.8.0"
 
 [workspace.dependencies.bottlerocket-settings-plugin]
 git = "/~https://github.com/bottlerocket-os/bottlerocket-settings-sdk"
-tag = "bottlerocket-settings-models-v0.7.0"
+tag = "bottlerocket-settings-models-v0.8.0"
 version = "0.1.0"
 
 [workspace.dependencies.settings-extension-oci-defaults]
 git = "/~https://github.com/bottlerocket-os/bottlerocket-settings-sdk"
-tag = "bottlerocket-settings-models-v0.7.0"
+tag = "bottlerocket-settings-models-v0.8.0"
 version = "0.1.0"
 
 [profile.release]

sources/settings-migrations/v1.34.0/kubelet-device-plugins-mig-settings/Cargo.toml

@@ -0,0 +1,15 @@
+[package]
+name = "kubelet-device-plugins-mig-settings"
+version = "0.1.0"
+authors = ["Piyush Jena <jepiyush@amazon.com>"]
+license = "Apache-2.0 OR MIT"
+edition = "2021"
+publish = false
+# Don't rebuild crate just because of changes to README.
+exclude = ["README.md"]
+
+
+# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+
+[dependencies]
+migration-helpers.workspace = true

sources/settings-migrations/v1.34.0/kubelet-device-plugins-mig-settings/src/main.rs

@@ -0,0 +1,22 @@
+use migration_helpers::common_migrations::AddPrefixesMigration;
+use migration_helpers::{migrate, Result};
+use std::process;
+
+/// We added new settings for configuring the NVIDIA k8s device plugin.
+fn run() -> Result<()> {
+    migrate(AddPrefixesMigration(vec![
+        "settings.kubelet-device-plugins.nvidia.device-partitioning-strategy",
+        "settings.kubelet-device-plugins.nvidia.mig",
+        "configuration-files.nvidia-k8s-device-plugin-mig-conf",
+    ]))
+}
+
+// Returning a Result from main makes it print a Debug representation of the error, but with Snafu
+// we have nice Display representations of the error, so we wrap "main" (run) and print any error.
+// /~https://github.com/shepmaster/snafu/issues/110
+fn main() {
+    if let Err(e) = run() {
+        eprintln!("{}", e);
+        process::exit(1);
+    }
+}

sources/shared-defaults/nvidia-k8s-device-plugin.toml

@@ -4,6 +4,7 @@ restart-commands = ["/bin/systemctl try-reload-or-restart nvidia-k8s-device-plug
 configuration-files = [
     "nvidia-k8s-device-plugin-conf",
     "nvidia-k8s-device-plugin-exec-start-conf",
+    "nvidia-k8s-device-plugin-mig-conf"
 ]
 
 [configuration-files.nvidia-k8s-device-plugin-conf]
@@ -14,6 +15,10 @@ template-path = "/usr/share/templates/nvidia-k8s-device-plugin-conf"
 path = "/etc/systemd/system/nvidia-k8s-device-plugin.service.d/exec-start.conf"
 template-path = "/usr/share/templates/nvidia-k8s-device-plugin-exec-start-conf"
 
+[configuration-files.nvidia-k8s-device-plugin-mig-conf]
+path = "/etc/nvidia-migmanager/nvidia-migmanager.toml"
+template-path = "/usr/share/templates/nvidia-k8s-device-plugin-mig-conf"
+
 [metadata.settings.kubelet-device-plugins.nvidia]
 affected-services = ["nvidia-k8s-device-plugin"]
 
@@ -22,3 +27,4 @@ pass-device-specs = true
 device-id-strategy="index"
 device-list-strategy="volume-mounts"
 device-sharing-strategy="none"
+device-partitioning-strategy="none"