snapshot: do not restore on offline nodes

A node may be offline. Since we select only one node, we should ensure it is not offline. Signed-off-by: Moritz Wanzenböck <[email protected]>
piraeusdatastore · Oct 25, 2024 · 4e80adf · 4e80adf
1 parent 64ad184
commit 4e80adf
Show file tree

Hide file tree

Showing 2 changed files with 15 additions and 2 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -10,6 +10,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 - Fixed an issue that caused the node publish call to attempt to resize a block volume: block volumes are now never
   attempted to be resized.
+- Ensure that snapshot restore requests are tried on nodes that are actually available.
 
 ## [1.6.3] - 2024-06-28
 

diff --git a/pkg/client/linstor.go b/pkg/client/linstor.go
@@ -1317,15 +1317,27 @@ func (s *Linstor) reconcileSnapshotResources(ctx context.Context, snapshot *volu
 
 	// Optimize the node we use to restore. It should be one of the preferred nodes, or just the first with a snapshot
 	// if no preferred nodes match.
-	selectedNode := snap.Nodes[0]
-
+	var selectedNode string
 	for _, snapNode := range snap.Nodes {
+		if err := s.NodeAvailable(ctx, snapNode); err != nil {
+			logger.WithField("selected node candidate", snapNode).WithError(err).Debug("node is not available")
+			continue
+		}
+
 		if slice.ContainsString(preferredNodes, snapNode) {
+			// We found a perfect candidate.
 			selectedNode = snapNode
 			break
+		} else if selectedNode == "" {
+			// Set a fallback if we have no candidate yet.
+			selectedNode = snapNode
 		}
 	}
 
+	if selectedNode == "" {
+		return fmt.Errorf("no node [%s] available for snapshot '%s'", strings.Join(snap.Nodes, ","), snap.Name)
+	}
+
 	logger.WithField("selected node", selectedNode).Debug("restoring snapshot on one node")
 
 	restoreConf := lapi.SnapshotRestore{ToResource: targetRD, Nodes: []string{selectedNode}}