Red Hat Bugzilla – Bug 905123
finishing storage migration offline / persistent dirty bitmap support
Last modified: 2017-04-27 16:14:16 EDT
When issuing the drive-mirror command for storage migration, you must restart the process from scratch if one of the following things happen: 1) QEMU crashes; 2) the VM is exited before QEMU is ready to pivot to the destination storage (before QEMU has reached "steady state").
This can be handled with a persistent dirty bitmap. A new instance of QEMU, or a new qemu-img command can then be used to complete storage migration.
Will not land in QEMU 2.4 and thus will not land in RHEL 7.2. Actively being worked on upstream, Expected for 2.5.
Still pending upstream as of 2.10 development window and will not be ready in time for proper testing in 7.4, unfortunately.
Bitmap persistence itself is a work in progress, but there are currently no discussions or plans for how precisely to implement offline storage migration, but it will happen almost certainly after persistence itself is completed.
I'll refrain from giving estimates... but please see upstream feature pages for updates: http://wiki.qemu.org/Features/IncrementalBackup