fix(gpu): handle scheduler inconsistency and device stuck in unhealthy

2026-01-05 16:39:56 +08:00
2 changed files with 2 additions and 2 deletions
--- a/infrastructure/gpu/.olares/config/gpu/hami/values.yaml
+++ b/infrastructure/gpu/.olares/config/gpu/hami/values.yaml
@@ -4,7 +4,7 @@ nameOverride: ""
 fullnameOverride: ""
 namespaceOverride: ""
 imagePullSecrets: []
-version: "v2.6.6"
+version: "v2.6.7"

 # Nvidia GPU Parameters
 resourceName: "nvidia.com/gpu"
--- a/platform/hami/.olares/Olares.yaml
+++ b/platform/hami/.olares/Olares.yaml
@@ -3,7 +3,7 @@ target: prebuilt
 output:
  containers:
    - 
-      name: beclab/hami:v2.6.6
+      name: beclab/hami:v2.6.7
    - 
      name: beclab/hami-webui-fe-oss:v1.0.8
    -