fix: prevent E2E test artifact leaks on TrueNAS

Adoption tests were silently ignoring cleanup errors (using `_ =`),
causing NVMe-oF subsystems, iSCSI targets/extents, and NFS shares to
accumulate on TrueNAS across CI runs.

- NVMe-oF/iSCSI adoption test 1: remove manual TrueNAS cleanup and let
  framework PVC cleanup trigger CSI DeleteVolume for proper cleanup
- All "markAdoptable" tests: replace `_ =` with Expect error checking
- iSCSI retain test: add missing target/extent cleanup before ZVOL
  deletion (NVMe-oF and NFS retain tests already did this)

Author: fenio

tests/e2e/iscsi/adoption_test.go

@@ -214,17 +214,9 @@ var _ = Describe("iSCSI Volume Adoption", func() {
 		err = f.K8s.WaitForPodDeleted(ctx, adoptedPodName, deleteTimeout)
 		Expect(err).NotTo(HaveOccurred())
 
-		By("Cleaning up the orphaned ZVOL from TrueNAS")
-		// Delete the iSCSI resources first if they were recreated
-		_ = f.TrueNAS.DeleteISCSITarget(ctx, newVolumeHandle)
-		_ = f.TrueNAS.DeleteISCSIExtent(ctx, newVolumeHandle)
-		_ = f.TrueNAS.DeleteISCSITarget(ctx, volumeHandle)
-		_ = f.TrueNAS.DeleteISCSIExtent(ctx, volumeHandle)
-		err = f.TrueNAS.DeleteDataset(ctx, zvolPath)
-		Expect(err).NotTo(HaveOccurred())
-		if f.Verbose() {
-			GinkgoWriter.Printf("Cleaned up orphaned ZVOL: %s\n", zvolPath)
-		}
+		// TrueNAS resources (iSCSI target, extent, ZVOL) are cleaned up
+		// by the framework's PVC cleanup, which triggers CSI DeleteVolume.
+		// The adopting StorageClass has no deleteStrategy=retain, so CSI performs full cleanup.
 	})
 
 	It("should mark a volume as adoptable when markAdoptable=true", func() {
@@ -296,8 +288,10 @@ var _ = Describe("iSCSI Volume Adoption", func() {
 		Expect(exists).To(BeTrue(), "Dataset should be retained with deleteStrategy=retain")
 
 		By("Cleaning up retained resources from TrueNAS")
-		_ = f.TrueNAS.DeleteISCSITarget(ctx, volumeHandle)
-		_ = f.TrueNAS.DeleteISCSIExtent(ctx, volumeHandle)
+		err = f.TrueNAS.DeleteISCSITarget(ctx, volumeHandle)
+		Expect(err).NotTo(HaveOccurred(), "Failed to delete retained iSCSI target from TrueNAS")
+		err = f.TrueNAS.DeleteISCSIExtent(ctx, volumeHandle)
+		Expect(err).NotTo(HaveOccurred(), "Failed to delete retained iSCSI extent from TrueNAS")
 		err = f.TrueNAS.DeleteDataset(ctx, zvolPath)
 		Expect(err).NotTo(HaveOccurred())
 		if f.Verbose() {

tests/e2e/iscsi/delete_strategy_retain_test.go

@@ -130,6 +130,14 @@ var _ = Describe("iSCSI Delete Strategy Retain", func() {
 		By("ZVOL confirmed to still exist on TrueNAS - retain strategy working correctly")
 		GinkgoWriter.Printf("Successfully verified ZVOL %s was retained on TrueNAS\n", zvolPath)
 
+		By("Cleaning up retained iSCSI target from TrueNAS")
+		err = f.TrueNAS.DeleteISCSITarget(ctx, volumeHandle)
+		Expect(err).NotTo(HaveOccurred(), "Failed to delete retained iSCSI target from TrueNAS")
+
+		By("Cleaning up retained iSCSI extent from TrueNAS")
+		err = f.TrueNAS.DeleteISCSIExtent(ctx, volumeHandle)
+		Expect(err).NotTo(HaveOccurred(), "Failed to delete retained iSCSI extent from TrueNAS")
+
 		By("Cleaning up retained ZVOL from TrueNAS")
 		err = f.TrueNAS.DeleteDataset(ctx, zvolPath)
 		Expect(err).NotTo(HaveOccurred(), "Failed to delete retained ZVOL from TrueNAS")
@@ -139,7 +147,7 @@ var _ = Describe("iSCSI Delete Strategy Retain", func() {
 		Expect(err).NotTo(HaveOccurred())
 		Expect(exists).To(BeFalse(), "ZVOL should no longer exist on TrueNAS after cleanup")
 
-		By("Cleanup verified - ZVOL successfully removed from TrueNAS")
+		By("Cleanup verified - ZVOL and iSCSI resources successfully removed from TrueNAS")
 		GinkgoWriter.Printf("Successfully cleaned up ZVOL %s from TrueNAS\n", zvolPath)
 	})
 })

tests/e2e/nfs/adoption_test.go

@@ -379,7 +379,8 @@ var _ = Describe("NFS Volume Adoption", func() {
 
 		By("Cleaning up retained resources from TrueNAS")
 		nfsSharePath := "/mnt/" + volumeHandle
-		_ = f.TrueNAS.DeleteNFSShare(ctx, nfsSharePath) // May fail if already deleted
+		err = f.TrueNAS.DeleteNFSShare(ctx, nfsSharePath)
+		Expect(err).NotTo(HaveOccurred(), "Failed to delete retained NFS share from TrueNAS")
 		err = f.TrueNAS.DeleteDataset(ctx, datasetPath)
 		Expect(err).NotTo(HaveOccurred())
 		if f.Verbose() {

tests/e2e/nvmeof/adoption_test.go

@@ -215,16 +215,9 @@ var _ = Describe("NVMe-oF Volume Adoption", func() {
 		err = f.K8s.WaitForPodDeleted(ctx, adoptedPodName, deleteTimeout)
 		Expect(err).NotTo(HaveOccurred())
 
-		By("Cleaning up the orphaned ZVOL from TrueNAS")
-		// Delete the subsystem first if it was recreated
-		newSubsystemNQN := "nqn.2137.csi.tns:" + newVolumeHandle
-		_ = f.TrueNAS.DeleteNVMeOFSubsystem(ctx, newSubsystemNQN)
-		_ = f.TrueNAS.DeleteNVMeOFSubsystem(ctx, subsystemNQN)
-		err = f.TrueNAS.DeleteDataset(ctx, zvolPath)
-		Expect(err).NotTo(HaveOccurred())
-		if f.Verbose() {
-			GinkgoWriter.Printf("Cleaned up orphaned ZVOL: %s\n", zvolPath)
-		}
+		// TrueNAS resources (NVMe-oF subsystem, port bindings, ZVOL) are cleaned up
+		// by the framework's PVC cleanup, which triggers CSI DeleteVolume.
+		// The adopting StorageClass has no deleteStrategy=retain, so CSI performs full cleanup.
 	})
 
 	It("should mark a volume as adoptable when markAdoptable=true", func() {
@@ -298,7 +291,8 @@ var _ = Describe("NVMe-oF Volume Adoption", func() {
 		Expect(exists).To(BeTrue(), "Dataset should be retained with deleteStrategy=retain")
 
 		By("Cleaning up retained resources from TrueNAS")
-		_ = f.TrueNAS.DeleteNVMeOFSubsystem(ctx, subsystemNQN)
+		err = f.TrueNAS.DeleteNVMeOFSubsystem(ctx, subsystemNQN)
+		Expect(err).NotTo(HaveOccurred(), "Failed to delete retained NVMe-oF subsystem from TrueNAS")
 		err = f.TrueNAS.DeleteDataset(ctx, zvolPath)
 		Expect(err).NotTo(HaveOccurred())
 		if f.Verbose() {