update release

ofek · May 16, 2021 · b84836f · b84836f
1 parent 85f39d2
commit b84836f
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 6 deletions.
diff --git a/deploy/overlays/stable/kustomization.yaml b/deploy/overlays/stable/kustomization.yaml
@@ -4,4 +4,4 @@ bases:
 - ../../base
 images:
 - name: docker.io/ofekmeister/csi-gcs
-  newTag: v0.7.0
+  newTag: v0.8.0
diff --git a/docs/troubleshooting.md b/docs/troubleshooting.md
@@ -2,7 +2,7 @@
 
 -----
 
-## Warnings events from pods
+## Early warnings from pods
 
 Warnings, like the one below, can be seen from pods scheduled on newly started nodes.
 
@@ -12,11 +12,13 @@ MountVolume.MountDevice failed for volume "xxxx" : kubernetes.io/csi: attacher.M
 
 Those warnings are temporary and reflect that the driver is still starting. Kubernetes will retry until the driver is ready. The problem is often encountered in clusters with auto-scaler as nodes come and go.
 
-This is a known issue of kubernetes (see [kubernetes/issues/75890](https://github.com/kubernetes/kubernetes/issues/75890)).
+This is a known issue of kubernetes (see [kubernetes#75890](https://github.com/kubernetes/kubernetes/issues/75890)).
 
-A possible workaround is to taint all nodes running the csi-gcs driver like `<driver name>/driver-ready=false:NoSchedule` and use, as suggested in [kubernetes/issues/75890#issuecomment-725792993](https://github.com/kubernetes/kubernetes/issues/75890#issuecomment-725792993), a custom controller like [wish/nodetaint](https://github.com/wish/nodetaint) to remove the taint once the csi-gcs pod is ready.
+A possible workaround is to taint all nodes running the `csi-gcs` driver like `<driver name>/driver-ready=false:NoSchedule` and use, as suggested in [this comment](https://github.com/kubernetes/kubernetes/issues/75890#issuecomment-725792993), a custom controller like [wish/nodetaint](https://github.com/wish/nodetaint) to remove the taint once the `csi-gcs` pod is ready.
 
-This workaround will ensure pods are repelled from nodes until the csi-gcs driver is ready without interfering with other components like the cluster auto-scaler.
+This workaround will ensure pods are repelled from nodes until the `csi-gcs` driver is ready without interfering with other components like the cluster auto-scaler.
+
+By default, `<driver name>` is `gcs.csi.ofek.dev`.
 
 !!! warning
-    csi-gcs label the node with `<driver name>/driver-ready=true` to reflect its readiness state. It's possible to use a node selector to select nodes with a ready csi-gcs node driver. However, it doesn't work with clusters using cluster-autoscaler as the auto-scaler will never find a node with matching `<driver name>/driver-ready=true` label. 
+    The driver labels the node with `<driver name>/driver-ready=true` to reflect its readiness state. It's possible to use a node selector to select nodes with a ready `csi-gcs` node driver. However, it doesn't work with clusters using cluster-autoscaler as the auto-scaler will never find a node with matching `<driver name>/driver-ready=true` label.