roachtest: don't reuse clusters that call dmsetup

Certain tests need to modify the blockdevice and they are prone to failures during setup that the device is still busy. Ideally we would figure out what is still holding onto the dish handle, but it is safer to simply not reuse clusters that perform this by adding `spec.ReuseNone()` Fixes: cockroachdb#107865 Epic: none Release note: None
rytaft · Aug 4, 2023 · 704e6e9 · 704e6e9
1 parent 9355e10
commit 704e6e9
Showing 1 changed file with 11 additions and 6 deletions.
diff --git a/pkg/cmd/roachtest/tests/failover.go b/pkg/cmd/roachtest/tests/failover.go
@@ -80,7 +80,7 @@ func registerFailover(r registry.Registry) {
 				Owner:               registry.OwnerKV,
 				Benchmark:           true,
 				Timeout:             60 * time.Minute,
-				Cluster:             r.MakeClusterSpec(10, spec.CPU(2), spec.PreferLocalSSD(false)), // uses disk stalls
+				Cluster:             r.MakeClusterSpec(10, spec.CPU(2), spec.PreferLocalSSD(false), spec.ReuseNone()), // uses disk stalls
 				Leases:              leases,
 				SkipPostValidations: registry.PostValidationNoDeadNodes, // cleanup kills nodes
 				Run: func(ctx context.Context, t test.Test, c cluster.Cluster) {
@@ -122,12 +122,17 @@ func registerFailover(r registry.Registry) {
 		for _, failureMode := range allFailureModes {
 			failureMode := failureMode // pin loop variable
 
-			var usePD bool
+			clusterOpts := make([]spec.Option, 0)
+			clusterOpts = append(clusterOpts, spec.CPU(2))
+
 			var postValidation registry.PostValidation
 			if failureMode == failureModeDiskStall {
 				// Use PDs in an attempt to work around flakes encountered when using
 				// SSDs. See #97968.
-				usePD = true
+				clusterOpts = append(clusterOpts, spec.PreferLocalSSD(false))
+				// Don't reuse the cluster for tests that call dmsetup to avoid
+				// spurious flakes from previous runs. See #107865
+				clusterOpts = append(clusterOpts, spec.ReuseNone())
 				postValidation = registry.PostValidationNoDeadNodes
 			}
 			r.Add(registry.TestSpec{
@@ -136,7 +141,7 @@ func registerFailover(r registry.Registry) {
 				Benchmark:           true,
 				Timeout:             30 * time.Minute,
 				SkipPostValidations: postValidation,
-				Cluster:             r.MakeClusterSpec(7, spec.CPU(2), spec.PreferLocalSSD(!usePD)),
+				Cluster:             r.MakeClusterSpec(7, clusterOpts...),
 				Leases:              leases,
 				Run: func(ctx context.Context, t test.Test, c cluster.Cluster) {
 					runFailoverNonSystem(ctx, t, c, failureMode)
@@ -149,7 +154,7 @@ func registerFailover(r registry.Registry) {
 				Benchmark:           true,
 				Timeout:             30 * time.Minute,
 				SkipPostValidations: postValidation,
-				Cluster:             r.MakeClusterSpec(5, spec.CPU(2), spec.PreferLocalSSD(!usePD)),
+				Cluster:             r.MakeClusterSpec(5, clusterOpts...),
 				Leases:              leases,
 				Run: func(ctx context.Context, t test.Test, c cluster.Cluster) {
 					runFailoverLiveness(ctx, t, c, failureMode)
@@ -162,7 +167,7 @@ func registerFailover(r registry.Registry) {
 				Benchmark:           true,
 				Timeout:             30 * time.Minute,
 				SkipPostValidations: postValidation,
-				Cluster:             r.MakeClusterSpec(7, spec.CPU(2), spec.PreferLocalSSD(!usePD)),
+				Cluster:             r.MakeClusterSpec(7, clusterOpts...),
 				Leases:              leases,
 				Run: func(ctx context.Context, t test.Test, c cluster.Cluster) {
 					runFailoverSystemNonLiveness(ctx, t, c, failureMode)