Skip to content

XSI-1969 XSI-1973 use https for localhost migration #6626

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: master
Choose a base branch
from

Conversation

lindig
Copy link
Contributor

@lindig lindig commented Aug 18, 2025

We observe an instability of localhost migration with NVidia SRIOV GPUs since we starting to fall back to http (rather than https) for localhost migration. Outside of development, localhost migration is used during device model restart.

Use HTTPS for localhost migration. This seems to improve stability (but is not the root cause).

We observe an instability of localhost migration with NVidia SRIOV GPUs
since we starting to fall back to http (rather than https) for localhost
migration. Outside of development, localhost migration is used during
device model restart.

Use HTTPS for localhost migration. This seems to improve stability (but
is not the root cause).

Signed-off-by: Christian Lindig <[email protected]>
@lindig lindig requested review from robhoes and edwintorok August 18, 2025 14:23
@robhoes
Copy link
Member

robhoes commented Aug 18, 2025

I can't see why HTTPS within a host would be useful, while it will affect performance. And this change will affect performance (and potentially stability) for many more use cases than vGPU localhost migration. So I am not in favour of this unless we really understand why it is better and there is no alternative.

@edwintorok
Copy link
Contributor

edwintorok commented Aug 18, 2025

Can we make it configurable via xapi.conf instead? That'd still allow us to test https on vs off to identify what the root cause is, but without having to bisect builds: we could do this test using the same (latest) build, by tweaking this parameter on/off.

@lindig
Copy link
Contributor Author

lindig commented Aug 18, 2025

This is purely on the observation that it appears to be more stable but it is not addressing a root problem. Localhost migration is not a typical use case but it failing is a problem - so I think we should do this if it improves the stability. @edwintorok the flag from xapi.conf still has an effect: migration-https-only = false forces to use HTTP.

@freddy77
Copy link
Collaborator

It looks like an indication there's a race condition doing the migration that goes away due to time changes.
Not saying the workaround is not worth.

@robhoes
Copy link
Member

robhoes commented Aug 18, 2025

so I think we should do this if it improves the stability

If I recall correctly, the change to not use HTTPS for localhost connections was also to improve stability. Also, this change affects all VMs, not just those with vGPUs

@lindig
Copy link
Contributor Author

lindig commented Aug 18, 2025

We could limit this to VMs with vGPU. But there is a cost: all these if-then-else mazes create untested areas.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants