Looking at the launchpad bug I see the original complaint was the disk creation was slow. This is somewhat surprising to me. falloc should be fast on any mainstream, modern filesystem as it merely has to write metadata in the FS to reserve the new blocks. I don't see any report of what host filesystem was in use when the user saw > 1 minute disk allocation. It would be worth investigating that aspect more, because my expectation is that falloc should be negligible in time.
Looking at the launchpad bug I see the original complaint was the disk creation was slow. This is somewhat surprising to me. falloc should be fast on any mainstream, modern filesystem as it merely has to write metadata in the FS to reserve the new blocks. I don't see any report of what host filesystem was in use when the user saw > 1 minute disk allocation. It would be worth investigating that aspect more, because my expectation is that falloc should be negligible in time.