Handle freeze on paused VMs during snapshot #15001

noamasu · 2025-06-25T07:51:00Z

VM snapshot with qemu-guest-agent fails when the domain is paused, because virt-controller sends a freeze/unfreeze request during snapshotting. This freeze/unfreeze is handled by virt-launcher, but it currently fails on paused VMs. Since paused VMs have a consistent filesystem state (similar to frozen), this change makes virt-launcher treat freeze as successful when the domain is paused. This allows snapshotting to continue safely.

What this PR does

Before this PR:

VM snapshot with qemu-guest-agent fails when the domain is paused. This happens since the snapshot flow trigger a freeze but since the VM is paused, freeze fails causing snapshot flow to also fail. since a Paused VM is considered a safe state to perform a snapshot, we expect it to pass when the VM is paused, but it doesnt.
it is possible to unpause a VM during snapshot, which is not desired.
a false warning message is being printed saying it is not safe to snapshot a running VM (without qemu-guest-agent) even though the VM is paused.
virt-launcher's agent poller is spamming qemu-gest-agent command failures when VM is paused.

After this PR:

Possible to complete a VM (with qemu-guest-agent) snapshot successfully even if the VM is paused.
Make sure unpause will be rejected if vm.Status.SnapshotInProgress is not nil.
prevent a warning message from being printed for when snapshot is being taken while VM (without qemu-guest-agent) is still running even though the VM is paused.
virt-launcher's agent poller skips qemu-guest-agent commands if the VM is paused.

References

Fixes #10759

Why we need it and why it was done in this way

The following tradeoffs were made:

The following alternatives were considered:

Links to places where the discussion took place:

Special notes for your reviewer

@ShellyKa13
@codingben (check out the change for agent_poller.go)

Checklist

This checklist is not enforcing, but it's a reminder of items that could be relevant to every PR.
Approvers are expected to review this list.

Design: A design document was considered and is present (link) or not required
PR: The PR description is expressive enough and will help future contributors
Code: Write code that humans can understand and Keep it simple
Refactor: You have left the code cleaner than you found it (Boy Scout Rule)
Upgrade: Impact of this change on upgrade flows was considered and addressed if required
Testing: New code requires new unit tests. New features and bug fixes require at least one e2e test
Documentation: A user-guide update was considered and is present (link) or not required. You want a user-guide update if it's a user facing feature / API change.
Community: Announcement to kubevirt-dev was considered

Release note

bugfix: Enable vmsnapshot for paused VMs

kubevirt-bot · 2025-06-25T07:51:10Z

Hi @noamasu. Thanks for your PR.

PRs from untrusted users cannot be marked as trusted with /ok-to-test in this repo meaning untrusted PR authors can never trigger tests themselves. Collaborators can still trigger tests on the PR using /test all.

I understand the commands that are listed here.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

sourcery-ai

Hey @noamasu - I've reviewed your changes and they look great!

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

0xFelix

Thanks for working on this, I've added some comments.

0xFelix · 2025-06-26T09:16:20Z

pkg/virt-launcher/virtwrap/manager.go

+		} else {
+			return err
+		}


Suggested change

} else {

return err

}

}

return err

nit: additional else not required

why do we need to check for not found and do a new error? why not return the error which is not found anyways?

0xFelix · 2025-06-26T09:17:02Z

pkg/virt-launcher/virtwrap/manager.go

+		} else {
+			return err
+		}


0xFelix · 2025-06-26T09:18:46Z

tests/storage/snapshot.go

+					} else {
+						vmi = libvmifact.NewAlpine(libnet.WithMasqueradeNetworking())
+					}
+					vmi.Namespace = testsuite.GetTestNamespace(nil)


There's also libvmi.WithNamespace. Can you put both libvmi Options into an opts slice?

0xFelix · 2025-06-26T09:19:28Z

tests/storage/snapshot.go

+					vm.Spec.Template.Spec.Domain.Devices.Disks = append(vm.Spec.Template.Spec.Domain.Devices.Disks, v1.Disk{
+						Name: "blank",
+						DiskDevice: v1.DiskDevice{
+							Disk: &v1.DiskTarget{
+								Bus: v1.DiskBusVirtio,
+							},
+						},
+					})
+					vm.Spec.Template.Spec.Volumes = append(vm.Spec.Template.Spec.Volumes, v1.Volume{
+						Name: "blank",
+						VolumeSource: v1.VolumeSource{
+							DataVolume: &v1.DataVolumeSource{
+								Name: "dv-" + vm.Name,
+							},
+						},
+					})


Add that DV to the vmi first (with libvmi) and then create a VM from it?

0xFelix · 2025-06-26T09:19:44Z

tests/storage/snapshot.go

+						},
+					})

+					vm = libvmi.NewVirtualMachine(vmi)


You call this func twice, also in L577

0xFelix · 2025-06-26T09:20:36Z

tests/storage/snapshot.go

+				Entry("[test_id:7000] with guest-agent", true),
+				Entry("[test_id:7001] no guest-agent", false),


Provide the libvmifact function here instead of a boolean?

@0xFelix , i tried to do:

Entry("[test_id:7000] with guest-agent", libvmifact.NewFedora), Entry("[test_id:7001] without guest-agent", libvmifact.NewAlpine),

insead of boolean,
but becuase i need this condition to wait for the guest agent:

if imageFunc == libvmifact.NewFedora { Eventually(matcher.ThisVMI(vmi), 12*time.Minute, 2*time.Second). Should(matcher.HaveConditionTrue(v1.VirtualMachineInstanceAgentConnected)) }

I cannot compare a function like imageFunc == libvmifact.NewFedora ... so seems like i have to pass a boolean anyways?
wondering what how to handle this if statement if i dont have the boolean

0xFelix · 2025-06-26T09:21:35Z

pkg/virt-launcher/virtwrap/agent-poller/agent_poller.go

 // GET_AGENT - According to libvirt engineers this command shouldn't be used
 // by KubeVirt, because it provides irrelevant information (version and supported commands).
 func executeAgentCommands(commands []AgentCommand, agentPoller *AgentPoller) {
+	dom, err := agentPoller.Connection.LookupDomainByName(agentPoller.domainName)


Is it correct that these changes are for virt-launcher's agent poller is spamming qemu-gest-agent command failures when VM is paused.? If so, can you put them into a separate commit please?

Is there a way to test this behavior?

Yes, i will make a separate commit for this, and add relevant tests to it :)

nice addition to the PR, thanks

ShellyKa13

Looks good! a couple of comments, also I will appreciate if you divide the PR to several commit
1.snapshot source change
2. manager freeze unfreeze changes + ut
3. lifecycle change + ut
4. extra change for running agent commands + ut
5. functional test

ShellyKa13 · 2025-06-26T14:37:39Z

pkg/virt-api/rest/lifecycle.go


 }

 func (app *SubresourceAPIApp) UnpauseVMIRequestHandler(request *restful.Request, response *restful.Response) {


should add this case to the subresource unit test

dont we also need to prevent from doing pause during snapshot? it might fail but should we even pass it along if it will just fail?

more I think about it I'm a bit conflicted on doing anything special in pause/unpause subresource since the VMI is not just paused/unpaused by users may happen when internal error occurs like disk i/o error

@mhenriks I think we should at least handle cases of unpause done by users..

Yeah I guess unpause is fine

ShellyKa13 · 2025-06-26T14:39:16Z

pkg/virt-api/rest/lifecycle.go

+		writeError(statusErr, response)
+		return
+	}
+	if vm.Status.SnapshotInProgress != nil {


I wonder if we want to add such a check in every state change subresource, I assume it is possible that the subresource will be used directly without modifing the vm yaml (where we deny with the admitter), @mhenriks wdyt?

hmm as far as I know, the subresources usually update something in the status and then a controller applies the change to the spec. When that happen the webhook will be invoked, no?

I verified with Cursor :) and as I thought in this cases the subresource call itself doesnt change vm/vmi which doesnt call the webhook and then the subresource passes the request directly to virt-handler which passes it right away to trough the grpc connection to the virt-handler to libvirt..

far be it from me to question the AI but I'm pretty sure not "every state change subresource" is processed like that. That may be the case for pause because that is fundamentally an operation on the VMI

@mhenriks technically any vm state change can come from virtctl(and I would assume UI also) bypassing the vm/vmi yaml change directly using the subresource. I think in such case the cases mentioned in the commit should be avoided is possible. We already do that in the webhook level so why not also subresource level?

Can you be more specific about what subresource calls? I know that VM start/stop subresource updates the VM spec/status which should invoke the validating webhook

right I see you are right, they call the subresource directly and the subresource patches the vm with start and stop. So I guess only pause/unpause does not go through the yaml

ShellyKa13 · 2025-06-26T14:40:51Z

pkg/virt-launcher/virtwrap/agent-poller/agent_poller.go

 // GET_AGENT - According to libvirt engineers this command shouldn't be used
 // by KubeVirt, because it provides irrelevant information (version and supported commands).
 func executeAgentCommands(commands []AgentCommand, agentPoller *AgentPoller) {
+	dom, err := agentPoller.Connection.LookupDomainByName(agentPoller.domainName)


nice addition to the PR, thanks

ShellyKa13 · 2025-06-26T14:41:28Z

pkg/virt-launcher/virtwrap/agent-poller/agent_poller.go

+	if paused, err := util.DomainIsPaused(dom); err != nil {
+		log.Log.Errorf("cannot determine domain state: %v", err)
+		return
+	} else if paused {


no need for the else
just do if paused

ShellyKa13 · 2025-06-26T14:43:00Z

pkg/virt-launcher/virtwrap/manager.go

+		} else {
+			return err
+		}


why do we need to check for not found and do a new error? why not return the error which is not found anyways?

ShellyKa13 · 2025-06-26T14:46:40Z

pkg/virt-launcher/virtwrap/manager.go

+	}
+	defer domain.Free()
+
+	if paused, err := util.DomainIsPaused(domain); err != nil {


since you added this function you can also use it in other places which check if domain is pause like:
UnpauseVMI

ShellyKa13 · 2025-06-26T14:50:00Z

tests/storage/snapshot.go

+				Entry("[test_id:7000] with guest-agent", true),
+				Entry("[test_id:7001] no guest-agent", false),


ShellyKa13 · 2025-06-26T14:51:58Z

pkg/storage/snapshot/source.go


 	online := exists

+	running := exists && s.vm.Status.PrintableStatus == kubevirtv1.VirtualMachineStatusRunning


@mhenriks what do you think about using PrintableStatus? its not being relied upon to a big thing just whether to print on not the error message, but usually we prefer not to rely on it

It's probably okay for Running but I wouldn't rely on it for anything else

maybe it would be best to just check the VMI conditions instead

codingben · 2025-06-29T11:24:37Z

pkg/virt-launcher/virtwrap/agent-poller/agent_poller.go

+	}
+	defer dom.Free()
+
+	if paused, err := util.DomainIsPaused(dom); err != nil {


Why did you added these changes only to executeAgentCommands()? because of the relevant QEMU commands in this function?

thanks for the response!
i added it because the agent poller's executeAgentCommands spam the logs with failed qemu agent command failures when the VM is paused (qemu guest agent cannot execute commands if the VM is not running). do you think we need to add it to other places?

There is also fetchAndStoreGuestInfo() that is using libvirt API to get relevant info instead of QEMU directly as an abstraction above it.

It seems that since fetchAndStoreGuestInfo() doesn't rely on QEMU guest agent commands, so it's acceptable to retrieve information via the libvirt API even when the VM is paused.

actually, i take it back, you are right.
domain.GetGuestInfo(infoTypes, 0) will also fail when the VM is paused, as these messages are also being spammed

{"component":"virt-launcher","level":"info","msg":"Polling API operations: 1","pos":"agent_poller.go:417","timestamp":"2025-06-29T18:45:50.316563Z"} {"component":"virt-launcher","level":"error","msg":"Fetching guest info failed: virError(Code=55, Domain=10, Message='Requested operation is not valid: domain is not running')","pos":"agent_poller.go:432","times

i will work on a fix that will cover both fetchAndStoreGuestInfo and executeAgentCommands
sorry for the confusion :)

mhenriks · 2025-07-01T03:02:51Z

Since paused VMs have a consistent filesystem state (similar to frozen),

@noamasu can you supply some documentation for this? I don't think buffers are flushed when a VM is paused. Consider that a VM may be paused automatically by an I/O error

mhenriks · 2025-07-01T03:06:36Z

pkg/storage/snapshot/source.go


 	s.state = &sourceState{
 		online:     online,
+		running:    running,


can we just have online be false if the VM is paused and not add a new state?

asked another way, why can't we simply treat a paused VM as offline?

noamasu · 2025-07-02T08:42:25Z

Hey @mhenriks, thank you for the feedback.

Since paused VMs have a consistent filesystem state (similar to frozen),
the assumption was according to the Jira ticket's expected result - "VMSnapshot completed skipping freeze if VM is paused, and not fail".

after looking into this - sorry for this inaccuracy, you are right, pause will keep memory unflushed - risking a non-stable snapshot of the filesystem.

looks like just flushing the suspended vm is also not possible since you need a running VM to do so.

indeed snapshot of a paused VM is not safe, much like a running VM (memory not flushed to disk)...

knowing that's the case what do you think is the best way to deal with this as I see 2 options here:

treat "paused" the same as "running"... which means to check if VM is paused/running (online) and print the existing warning message of "not safe to snapshot a paused/running VM". do you think a quesedFailed indication should be presented in that case?
in the freeze function (snapshot/source.go): if the vm is paused - unpause the vm before freeze, then do the freeze -> snapshot -> unfreeze regularly, then pause the vm again. (I dont think its ok to change the the VM state tbh)

mhenriks · 2025-07-02T20:28:13Z

treat "paused" the same as "running"... which means to check if VM is paused/running (online) and print the existing warning message of "not safe to snapshot a paused/running VM". do you think a quesedFailed indication should be presented in that case?

@noamasu I think this is the better option, may even want to create a new indication for this case

mhenriks · 2025-07-08T14:59:22Z

I think this is looking pretty good, what are your thoughts @ShellyKa13?

ShellyKa13

Couple of suggestions

ShellyKa13 · 2025-07-10T08:42:15Z

pkg/storage/snapshot/snapshot.go

 		}
+
+		// Check for paused VM
+		if source.Paused() {


What do you say about checking this before the guestagent and in case of paused not add the guestagent indication since it doesnt affect the snapshot. i.e
if source.Paused{
...
else if source.GuestAgent{
...
} else{
...
}

ShellyKa13 · 2025-07-10T08:44:22Z

pkg/storage/snapshot/source.go

 		return nil
 	}

+	if s.Paused() {


why not add it to the first if:
if !s.Locked() || !s.GuestAgent() || s.Paused() {

ShellyKa13 · 2025-07-10T08:45:20Z

staging/src/kubevirt.io/api/snapshot/v1beta1/types.go

 	VMSnapshotNoGuestAgentIndication   Indication = "NoGuestAgent"
 	VMSnapshotGuestAgentIndication     Indication = "GuestAgent"
 	VMSnapshotQuiesceFailedIndication  Indication = "QuiesceFailed"
+	VMSnapshotPausedIndication         Indication = "Paused"


This should probable be in the previous commit, or move the indications part from previous commit to this commit

ShellyKa13 · 2025-07-10T09:05:43Z

pkg/virt-api/rest/lifecycle.go

+		writeError(statusErr, response)
+		return
+	}
+	if vm.Status.SnapshotInProgress != nil {


I verified with Cursor :) and as I thought in this cases the subresource call itself doesnt change vm/vmi which doesnt call the webhook and then the subresource passes the request directly to virt-handler which passes it right away to trough the grpc connection to the virt-handler to libvirt..

ShellyKa13 · 2025-07-10T09:08:13Z

pkg/virt-api/rest/lifecycle.go


 }

 func (app *SubresourceAPIApp) UnpauseVMIRequestHandler(request *restful.Request, response *restful.Response) {


dont we also need to prevent from doing pause during snapshot? it might fail but should we even pass it along if it will just fail?

ShellyKa13 · 2025-07-15T14:29:21Z

pkg/storage/snapshot/source.go

 		return nil
 	}

+	if s.Paused() {


probably should first check the paused condition before the guestagent one

ShellyKa13 · 2025-07-15T14:33:19Z

tests/storage/snapshot.go

+								return updatedVMI.Status.FSFreezeStatus
+		}, 30*time.Second, 2*time.Second).Should(BeEmpty())
+	})
+	DescribeTable("should succeed snapshot when VM is paused with Paused indication",


missing new line before this block

ShellyKa13 · 2025-07-15T14:35:14Z

tests/storage/snapshot.go

+			checkOnlineSnapshotExpectedContentSource(vm, contentName, true)
+		},
+
+		Entry("[test_id:7000] with guest-agent", libvmifact.NewFedora, true),


not sure if we really need to run this test twice for with or without guest agent.. I think with guest agent is enough

ShellyKa13 · 2025-07-15T14:35:29Z

pkg/storage/snapshot/snapshot.go


-		if source.GuestAgent() {
+		if source.Paused() {
+			indications = sets.Insert(indications, snapshotv1.VMSnapshotPausedIndication)


what about a unit test for this?

ShellyKa13 · 2025-07-15T14:38:20Z

pkg/virt-api/rest/lifecycle.go


 }

 func (app *SubresourceAPIApp) UnpauseVMIRequestHandler(request *restful.Request, response *restful.Response) {


@mhenriks I think we should at least handle cases of unpause done by users..

ShellyKa13 · 2025-07-15T14:40:10Z

pkg/virt-api/rest/lifecycle.go

+		writeError(statusErr, response)
+		return
+	}
+	if vm.Status.SnapshotInProgress != nil {


@mhenriks technically any vm state change can come from virtctl(and I would assume UI also) bypassing the vm/vmi yaml change directly using the subresource. I think in such case the cases mentioned in the commit should be avoided is possible. We already do that in the webhook level so why not also subresource level?

ShellyKa13 · 2025-07-15T14:42:37Z

pkg/virt-api/rest/lifecycle.go

+		writeError(statusErr, response)
+		return
+	}
+	if vm.Status.SnapshotInProgress != nil {


Dont think we can do this in the unfreeze since the snapshot is still in progress when you call unfreeze to complete the snapshot. And also there is an automatic unfreeze mechanism we use the prevent a case the vm is kept frozen if something happen to the snapshot and it for some reason doesnt unfreeze.

ShellyKa13 · 2025-07-15T14:44:22Z

pkg/virt-api/rest/lifecycle.go

 		}
 	}
-	_, err := app.fetchVirtualMachine(name, namespace)
+	vm, err := app.fetchVirtualMachine(name, namespace)


also in this case Im not sure we want to prevent migrate vm call.. I believe the snapshot will just fail if the vm was migrated in the process and if not dont think it will cause data change anyways

ShellyKa13 · 2025-07-27T07:30:07Z

/lgtm

kubevirt-bot · 2025-08-03T10:34:36Z

Pull requests that are marked with lgtm should receive a review
from an approver within 1 week.

After that period the bot marks them with the label needs-approver-review.

/label needs-approver-review

mhenriks · 2025-08-04T14:48:38Z

/approve

kubevirt-bot · 2025-08-04T14:48:51Z

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: mhenriks

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Details

Needs approval from an approver in each of these files:

~~OWNERS~~ [mhenriks]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

kubevirt-commenter-bot · 2025-08-04T14:48:55Z

Required labels detected, running phase 2 presubmits:
/test pull-kubevirt-e2e-k8s-1.31-windows2016
/test pull-kubevirt-e2e-kind-1.33-vgpu
/test pull-kubevirt-e2e-kind-sriov
/test pull-kubevirt-e2e-k8s-1.33-ipv6-sig-network
/test pull-kubevirt-e2e-k8s-1.31-sig-network
/test pull-kubevirt-e2e-k8s-1.31-sig-storage
/test pull-kubevirt-e2e-k8s-1.31-sig-compute
/test pull-kubevirt-e2e-k8s-1.31-sig-operator
/test pull-kubevirt-e2e-k8s-1.32-sig-network
/test pull-kubevirt-e2e-k8s-1.32-sig-storage
/test pull-kubevirt-e2e-k8s-1.32-sig-compute
/test pull-kubevirt-e2e-k8s-1.32-sig-operator

kubevirt-bot · 2025-08-05T01:59:40Z

/remove-label needs-approver-review

ShellyKa13 · 2025-08-05T06:18:46Z

I think we should backport it the relevant issue #10759 was opened several versions ago

noamasu · 2025-08-06T09:05:24Z

/cherry-pick release-1.6

kubevirt-bot · 2025-08-06T09:05:32Z

@noamasu: only kubevirt org members may request cherry picks. If you are already part of the org, make sure to change your membership to public. Otherwise you can still do the cherry-pick manually.

Details

In response to this:

/cherry-pick release-1.6

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

ShellyKa13 · 2025-08-06T09:08:08Z

/cherry-pick release-1.6

kubevirt-bot · 2025-08-06T09:08:48Z

@ShellyKa13: new pull request created: #15385

Details

In response to this:

/cherry-pick release-1.6

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

ShellyKa13 · 2025-08-06T09:09:43Z

/cherry-pick release-1.5

kubevirt-bot · 2025-08-06T09:10:27Z

@ShellyKa13: #15001 failed to apply on top of branch "release-1.5":

Applying: Treat paused VMs as unsafe for snapshots
Applying: Add Paused indication for VM snapshots
Using index info to reconstruct a base tree...
M	pkg/storage/snapshot/snapshot.go
M	pkg/storage/snapshot/snapshot_test.go
M	staging/src/kubevirt.io/api/snapshot/v1beta1/types.go
M	tests/storage/snapshot.go
Falling back to patching base and 3-way merge...
Auto-merging tests/storage/snapshot.go
CONFLICT (content): Merge conflict in tests/storage/snapshot.go
Auto-merging staging/src/kubevirt.io/api/snapshot/v1beta1/types.go
Auto-merging pkg/storage/snapshot/snapshot_test.go
Auto-merging pkg/storage/snapshot/snapshot.go
error: Failed to merge in the changes.
hint: Use 'git am --show-current-patch=diff' to see the failed patch
hint: When you have resolved this problem, run "git am --continue".
hint: If you prefer to skip this patch, run "git am --skip" instead.
hint: To restore the original branch and stop patching, run "git am --abort".
hint: Disable this message with "git config advice.mergeConflict false"
Patch failed at 0002 Add Paused indication for VM snapshots

Details

In response to this:

/cherry-pick release-1.5

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

Manual backport of kubevirt#15001 Validates vm.Status.SnapshotInProgress and refuses operations if it's non-nil, preventing VMs from being modified mid-snapshot. This ensures snapshot consistency by preventing state changes during the snapshot process, which could lead to undefined behavior or inconsistent snapshots. Operations protected: - UnpauseVMIRequestHandler Signed-off-by: Noam Assouline <[email protected]>

manual backport of kubevirt#15001 Introduces VMSnapshotPausedIndication to provide programmatic indication when snapshots are taken of paused VMs. This allows users and tooling to identify snapshots that may have consistency issues due to unflushed memory buffers. The indication appears in the VirtualMachineSnapshot status, consistent with existing indications like QuiesceFailed and NoGuestAgent. Includes functional tests to verify the indication is properly set. Signed-off-by: Noam Assouline <[email protected]> lala

Manual backport of kubevirt#15001 Validates vm.Status.SnapshotInProgress and refuses operations if it's non-nil, preventing VMs from being modified mid-snapshot. This ensures snapshot consistency by preventing state changes during the snapshot process, which could lead to undefined behavior or inconsistent snapshots. Operations protected: - UnpauseVMIRequestHandler Signed-off-by: Noam Assouline <[email protected]>

manual backport of kubevirt#15001 Introduces VMSnapshotPausedIndication to provide programmatic indication when snapshots are taken of paused VMs. This allows users and tooling to identify snapshots that may have consistency issues due to unflushed memory buffers. The indication appears in the VirtualMachineSnapshot status, consistent with existing indications like QuiesceFailed and NoGuestAgent. Includes functional tests to verify the indication is properly set. Signed-off-by: Noam Assouline <[email protected]>

Manual backport of kubevirt#15001 Validates vm.Status.SnapshotInProgress and refuses operations if it's non-nil, preventing VMs from being modified mid-snapshot. This ensures snapshot consistency by preventing state changes during the snapshot process, which could lead to undefined behavior or inconsistent snapshots. Operations protected: - UnpauseVMIRequestHandler Signed-off-by: Noam Assouline <[email protected]>

kubevirt-bot added dco-signoff: yes Indicates the PR's author has DCO signed all their commits. do-not-merge/release-note-label-needed Indicates that a PR should not merge because it's missing one of the release note labels. labels Jun 25, 2025

kubevirt-bot added the size/L label Jun 25, 2025

kubevirt-bot requested review from 0xFelix, ShellyKa13, alromeros and enp0s3 June 25, 2025 07:51

kubevirt-bot added area/launcher sig/compute sig/storage labels Jun 25, 2025

sourcery-ai bot reviewed Jun 25, 2025

View reviewed changes

0xFelix reviewed Jun 26, 2025

View reviewed changes

ShellyKa13 reviewed Jun 26, 2025

View reviewed changes

codingben reviewed Jun 29, 2025

View reviewed changes

noamasu force-pushed the bugfix/snapshot_when_vm_paused_fixes branch from c0cf59f to 680daf9 Compare June 30, 2025 13:30

mhenriks reviewed Jul 1, 2025

View reviewed changes

noamasu mentioned this pull request Jul 3, 2025

Stop agent poller when VM is paused #15071

Merged

8 tasks

noamasu force-pushed the bugfix/snapshot_when_vm_paused_fixes branch from 680daf9 to 44c254b Compare July 6, 2025 21:06

kubevirt-bot added the kind/api-change Categorizes issue or PR as related to adding, removing, or otherwise changing an API label Jul 6, 2025

noamasu force-pushed the bugfix/snapshot_when_vm_paused_fixes branch 2 times, most recently from ec76470 to d0da165 Compare July 6, 2025 21:45

ShellyKa13 reviewed Jul 10, 2025

View reviewed changes

noamasu force-pushed the bugfix/snapshot_when_vm_paused_fixes branch from 35e2247 to ef98e10 Compare July 15, 2025 13:32

ShellyKa13 reviewed Jul 15, 2025

View reviewed changes

noamasu force-pushed the bugfix/snapshot_when_vm_paused_fixes branch from ef98e10 to 4d6d98b Compare July 17, 2025 12:45

kubevirt-bot removed the lgtm Indicates that a PR is ready to be merged. label Jul 24, 2025

kubevirt-bot added the lgtm Indicates that a PR is ready to be merged. label Jul 27, 2025

kubevirt-bot added the needs-approver-review Indicates that a PR requires a review from an approver. label Aug 3, 2025

kubevirt-bot added the approved Indicates a PR has been approved by an approver from all required OWNERS files. label Aug 4, 2025

kubevirt-bot merged commit 1d23e32 into kubevirt:main Aug 4, 2025
41 checks passed

kubevirt-bot removed the needs-approver-review Indicates that a PR requires a review from an approver. label Aug 5, 2025

kubevirt-bot mentioned this pull request Aug 6, 2025

[release-1.6] Handle freeze on paused VMs during snapshot #15385

Merged

noamasu mentioned this pull request Sep 2, 2025

[Release-1.5] Handle freeze on paused VMs during snapshot #15568

Merged

8 tasks

noamasu mentioned this pull request Sep 9, 2025

[Release-1.4] Handle freeze on paused VMs during snapshot #15633

Merged

8 tasks

noamasu mentioned this pull request Sep 14, 2025

[Release-1.3] Handle freeze on paused VMs during snapshot #15664

Merged

8 tasks

noamasu mentioned this pull request Nov 24, 2025

Add noamasu to KubeVirt Members kubevirt/project-infra#4535

Merged

7 tasks

		Entry("[test_id:7000] with guest-agent", true),
		Entry("[test_id:7001] no guest-agent", false),


		}

		func (app SubresourceAPIApp) UnpauseVMIRequestHandler(request restful.Request, response *restful.Response) {


		online := exists

		running := exists && s.vm.Status.PrintableStatus == kubevirtv1.VirtualMachineStatusRunning

Handle freeze on paused VMs during snapshot #15001

Handle freeze on paused VMs during snapshot #15001

Uh oh!

Conversation

noamasu commented Jun 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What this PR does

Before this PR:

After this PR:

References

Why we need it and why it was done in this way

Special notes for your reviewer

Checklist

Release note

Uh oh!

kubevirt-bot commented Jun 25, 2025

Uh oh!

sourcery-ai bot left a comment

Choose a reason for hiding this comment

Uh oh!

0xFelix left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ShellyKa13 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

noamasu commented Jun 25, 2025 •

edited

Loading