fix nilpointer panic when decode addlitional item failed in restore

[AlbeeSo]

Thank you for contributing to Velero!

Please add a summary of your change

When decode additional item in restoreItem failed,

velero/pkg/restore/restore.go

Line 1390 in 78c97d9

errs.Add(namespace, errors.Wrapf(err, "error restoring additional item %s", additionalResourceID))

may case nilpointer panic in the inner function (like GetName, for exmple, as the obj is nil actually)

velero/pkg/restore/restore.go

Line 1400 in 78c97d9

w, e, additionalItemExists := ctx.restoreItem(additionalObj, additionalItem.GroupResource, additionalItemNamespace)

with messages like (I tested in velero 1.14, but the related logic is same with the latest):

panic: runtime error: invalid memory address or nil pointer dereference [recovered]
  panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x0 pc=0x241f664]

goroutine 350 [running]:
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Reconcile.func1()
  /go/pkg/mod/sigs.k8s.io/[email protected]/pkg/internal/controller/controller.go:116 +0x1e5
panic({0x281c160?, 0x4725ed0?})
  /usr/local/go/src/runtime/panic.go:770 +0x132
k8s.io/apimachinery/pkg/apis/meta/v1/unstructured.(*Unstructured).GetName(...)
  /go/pkg/mod/k8s.io/[email protected]/pkg/apis/meta/v1/unstructured/unstructured.go:251
github.com/vmware-tanzu/velero/pkg/restore.(*restoreContext).restoreItem(0xc0006a66c8, 0x0, {{0x0, 0x0}, {0xc0017a7ec0, 0x11}}, {0x0, 0x0})
  /go/src/github.com/vmware-tanzu/velero/pkg/restore/restore.go:1087 +0x1e4
github.com/vmware-tanzu/velero/pkg/restore.(*restoreContext).restoreItem(0xc0006a66c8, 0xc0004b1a28, {{0x0, 0x0}, {0xc0004e4120, 0x16}}, {0xc000af003c, 0x6})
  /go/src/github.com/vmware-tanzu/velero/pkg/restore/restore.go:1394 +0xb026
github.com/vmware-tanzu/velero/pkg/restore.(*restoreContext).processSelectedResource(0xc0006a66c8, {{0xc0004e4120?, 0xc0007cc7d0?}, 0xc001c8f4d0?, 0xc000c0e780?}, 0x377, 0x5c, 0xc001a96708, 0xc001182120)
  /go/src/github.com/vmware-tanzu/velero/pkg/restore/restore.go:757 +0xb34
github.com/vmware-tanzu/velero/pkg/restore.(*restoreContext).execute(0xc0006a66c8)
  /go/src/github.com/vmware-tanzu/velero/pkg/restore/restore.go:609 +0x31ca
github.com/vmware-tanzu/velero/pkg/restore.(*kubernetesRestorer).RestoreWithResolvers(0xc000685d00, 0xc000501340, {{0xc000ab4000?, 0xc00097b1c8?, 0xddeb5c?}}, {0x7f7be4018ff0, 0xc0002d2820})
  /go/src/github.com/vmware-tanzu/velero/pkg/restore/restore.go:326 +0x13d1
github.com/vmware-tanzu/velero/pkg/controller.(*restoreReconciler).runValidatedRestore(0xc0007c0240, 0xc00012d608, {0xc0007c6a88?, 0xc000acd6c0?}, 0x0)
  /go/src/github.com/vmware-tanzu/velero/pkg/controller/restore_controller.go:565 +0x1273
github.com/vmware-tanzu/velero/pkg/controller.(*restoreReconciler).Reconcile(0xc0007c0240, {0x3144e00, 0xc00095e900}, {{{0xc000968cc8?, 0x0?}, {0xc000058f60?, 0xc00095e900?}}})
  /go/src/github.com/vmware-tanzu/velero/pkg/controller/restore_controller.go:264 +0xb45
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Reconcile(0x314cfb8?, {0x3144e00?, 0xc00095e900?}, {{{0xc000968cc8?, 0xb?}, {0xc000058f60?, 0x0?}}})
  /go/pkg/mod/sigs.k8s.io/[email protected]/pkg/internal/controller/controller.go:119 +0xb7
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).reconcileHandler(0xc0005f6780, {0x3144e38, 0xc00033ce10}, {0x297d740, 0xc0004766a0})
  /go/pkg/mod/sigs.k8s.io/[email protected]/pkg/internal/controller/controller.go:316 +0x3bc
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).processNextWorkItem(0xc0005f6780, {0x3144e38, 0xc00033ce10})
  /go/pkg/mod/sigs.k8s.io/[email protected]/pkg/internal/controller/controller.go:266 +0x1be
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Start.func2.2()
  /go/pkg/mod/sigs.k8s.io/[email protected]/pkg/internal/controller/controller.go:227 +0x79
created by sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Start.func2 in goroutine 248
  /go/pkg/mod/sigs.k8s.io/[email protected]/pkg/internal/controller/controller.go:223 +0x50c

so I guess the right way is to give up restoring and just throw errors.

Does your change fix a particular issue?

Fixes #(issue)

Please indicate you've done the following:

• Accepted the DCO. Commits without the DCO will delay acceptance.
• Created a changelog file (make new-changelog) or comment /kind changelog-not-required on this PR.
• Updated the corresponding documentation in site/content/docs/main.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 59.18%. Comparing base (78c97d9) to head (24860e4).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #8563      +/-   ##
==========================================
+ Coverage   59.17%   59.18%   +0.01%     
==========================================
  Files         370      370              
  Lines       39487    39493       +6     
==========================================
+ Hits        23366    23375       +9     
+ Misses      14647    14645       -2     
+ Partials     1474     1473       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[AlbeeSo]

/kind changelog-not-required

[anshulahuja98]

what is the scenario where this was actually hit?
Shouldn't we fix the rootcause?
Or was it a internal plugin for some user

[AlbeeSo]

what is the scenario where this was actually hit? Shouldn't we fix the rootcause? Or was it a internal plugin for some user

hi, sorry that i am still trying reproduce the problem as the testing k8s cluster has been deleted.
but i think this PR is common for such cases. i will submit another PR if needed if i find something.

[anshulahuja98]

I am not sure if it is right to continue the restore by just putting this in error.
It should ideally be a hard failure.

[AlbeeSo]

I am not sure if it is right to continue the restore by just putting this in error. It should ideally be a hard failure.

A hard failure is another choice, instead of the panic currently.
However I still think putting error is enough, to keep consistent with the normal items
https://github.com/vmware-tanzu/velero/blob/03d0bd9d22940f56a2fd30bd1f66c17a9b8a6f30/pkg/restore/restore.go#L750C1-L761C5