(Windows) guest IPv6 address doesn't collapse zeroes -> Long IPv6 addresses

probain

I will check back tomorrow when I'm back at that lab-pool. Sorry for taking your time thus far. I will provide more and better info tomorrow

dinhngtu

This seems to be specific to XenServer guest agents?

jr-m4

@olivierlambert said:

Well, I'm not seeing the same behavior here, fe80::4496:9d87:690d:ccc7 is collapsed, unlike in your screenshot

I'm now back with better information from the original pool.

xe vm-list uuid=0513005e-3626-1867-306d-9d5415cd6516 params=networks
networks (MRO)    : 0/ip: 192.168.97.60; 0/ipv4/0: 192.168.97.60; 0/ipv6/0: fe80:0000:0000:0000:a57c:c544:3b03:8910

I will apply latest updates to see if this has any bearing on the matter.

jr-m4

After applying latest updates and commits, to both xcp and XO. The zeroes were still non-collapsed.

However. Rebooting the VM resulted in the zeroes now collapse properly as expected.

I had previously rebooted the VM, so I don't believe that was the solution. But I did not check inbetween pulling latest source, and updating XCP. So I'm unsure which update ~~solved it.~~

~~But this then has been resolved.~~
Update: Not solved! Now on fully updated XCP & XO, and the VM being rebooted as per post-details above.

dinhngtu

@jr-m4 Could you check if your VMs are using XenServer or XCP-ng tools?

jr-m4

@dinhngtu said:

@jr-m4 Could you check if your VMs are using XenServer or XCP-ng tools?

They are using the vates provided xcp-ng guest tools.

jr-m4

I have now tried updating to the latest version included in the xcp-ng guest-tools.iso. Followed by a reboot.

At the moment, the zeroes are collapsed. But I won't count my chickens before they've hatched. So I will reply back if they expand out again (as seen before).

Thanks!

jr-m4

I believe I have a reproducible method to make it expand.
The trick aparently is Live migrating the VM.

After reboot

Pre live migration:

Post live migration:

TeddyAstie

Do you have multiple guests agents in the VM (e.g Citrix and XCP-ng ones) that may step on each other for this IP ?

jr-m4

@TeddyAstie

I do not have Citrix ones. This VM was pretty much freshly installed to test Windows2025 on as pure Vates-maintained stack as possible.

Update: Please note my post above. Tha this seems to be triggered by live-migrating the vm

jr-m4

ping @poddingue for visibility as I believe I've found the trigger for IPv6 expanding... Live Migrations

poddingue

Thanks for the ping and for narrowing it down; that live-migration repro is a really useful signal.
I don't know enough about how the guest tools report IPs back through XAPI to say where the canonicalisation should happen, but it sounds like something @Team-Hypervisor-Kernel might want to look at since the trigger is on the agent side after migration.
If it turns out to be reproducible on another Windows guest version (2022, 2019), that might help narrow it further; no pressure though, you've already done the hard part.

jr-m4

@poddingue

I have done some more testing.
Windows Server 2022 is also affected. 9.1.145-77

I also tried the latest WindowsPV drivers from Vates, on Windows Server 2025. 9.1.200-0. And the "problem" still persists.

It is a very low impact and low importance "problem". Since it is afaik purely costmetical.

olivierlambert

Still, something to fix