vmm, vm-migration: add MigrationStateOngoingPhase::Started#178
Conversation
7d8796b to
ae2e4f2
Compare
There was a problem hiding this comment.
The commit referenced in the commit message for vmm, vm-migration: streamline error chain message with upstream should be cloud-hypervisor@2527020, since that is the one in main.
No action required, but just want to point out that I would have preferred to have two PRs, since the changes don't seem related.
ae2e4f2 to
be02ad7
Compare
thanks - I agree. I already did a split-out nevertheless to keep changes clean |
This will help us in libvirt/ch to distinguish short send/receive races from a real migration startup failure. By being able to check if the migration has been successfully started (rather than just the migration worker thread), we can improve the reliability of the start_migration() logic in libvirt/ch. On-behalf-of: SAP philipp.schuster@sap.com Signed-off-by: Philipp Schuster <philipp.schuster@cyberus-technology.de>
be02ad7 to
179eb5d
Compare
| // Signal that the migration connection has been established. Management | ||
| // software can use this to distinguish short send/receive races from a | ||
| // real migration startup failure. | ||
| { |
There was a problem hiding this comment.
btw @arctic-alpaca I know that this is rather poor design that we have to follow here - it comes from our "pragmatic" impl in the fork. Will (and must) be done much nicer when we upstream it
cloud-hypervisor: 2026-06-29T16:20:34+02:00 -> 2026-07-02T13:20:05+02:00 This includes the new migration phase marker [0] and [1]. [0] cyberus-technology/cloud-hypervisor#178 [1] cyberus-technology/cloud-hypervisor#179 This commit was generated via: nix run github:phip1611/nixos-configs#flake-update-and-commit On-behalf-of: SAP philipp.schuster@sap.com Signed-off-by: Philipp Schuster <philipp.schuster@cyberus-technology.de>
My proposed solution for https://github.com/cobaltcore-dev/cobaltcore/issues/598. Needs libvirt patches additionally (https://gitlab.cyberus-technology.de/cyberus/cloud/libvirt/-/merge_requests/251/).
This will help us in libvirt/ch to distinguish short send/receive races
from a real migration startup failure. By being able to check if the
migration has been successfully started (rather than just the migration
worker thread), we can improve the reliablity of the start_migration()
logic in libvirt/ch.
libvirt CI pipeline: https://gitlab.cyberus-technology.de/cyberus/cloud/libvirt/-/merge_requests/251/pipelines