This small patch series adds additional checks for VMs and CTs to avoid migrating local disk images multiple times.
The idea on how to reference them is similar as in the check if a disk should be added as unusedX [0]. For example, if there are multiple storage definitions pointing to the same actual storage, we can see the same image multiple times. These are ignored. If we notice that a disk image is configured multiple times in the config for the guest, we fail the migration with a warning. One thing I am not really sure, is the case where a disk image is currently configured fully and a second time as unusedX. At least the Qemu part won't let us remove it as it is still in use. But this is still a somewhat hypothetical example because in my tests, once we detach a double configured disk image (e.g. scsi0 and scsi1), it will not be added as unused0 but just disappear. So there should not be a reasonable way to end up with a unused disk config pointing to the same image as a configured one. [0] https://git.proxmox.com/?p=qemu-server.git;a=blob;f=PVE/QemuServer.pm;h=c1d0fd2d06a35bd8d996a59b44eb60345165f1b6;hb=HEAD#l6910 qemu-server: Aaron Lauterer (1): migration: avoid migrating disk images multiple times PVE/QemuMigrate.pm | 33 +++++++++++++++++++++++++++++++++ 1 file changed, 33 insertions(+) pve-container: Aaron Lauterer (1): migration: avoid migrating volume images multiple times src/PVE/LXC/Migrate.pm | 31 +++++++++++++++++++++++++++++++ 1 file changed, 31 insertions(+) -- 2.30.2 _______________________________________________ pve-devel mailing list pve-devel@lists.proxmox.com https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
