Created attachment 1979993 [details] The daemon log for the steps above Description of problem: As subject Version-Release number of selected component (if applicable): libvirt-9.5.0-3.el9.x86_64 qemu-kvm-8.0.0-9.el9.x86_64 How reproducible: 100% Steps to Reproduce: 1. Prepare two hosts for migration. Start a guest on one host 2. Migrate the VM with --copy-storage-all ➜ ~ virsh migrate --live rhel-9.2 qemu+ssh://root$DEST/system --copy-storage-all 3. At the same time of step2, try to cancel the migration by blockjob --abort ➜ ~ virsh blockjob rhel-9.2 vda --abort error: Requested operation is not valid: domain is not running Actual results: As above Expected results: Although "blockjob --abort" should not cancel the migration, it should return the operation unsupported err directly. Additional info:
BTW, the "blockjob --abort" sub-command is blocked until the migration is finished
Note that this behaviour is exactly the same as any other command which would modify the state of the VM, such as unplugging a device. The blockjob itself is considered internal for the use of the migration. In the above description I don't see any form of justification why we should treat the blockjobs differently than anything else, could you please elaborate?