I have a bunch of read-only snapshots that have an identical generation per the default output from 'btrfs sub list -t /mountpoint/' even though the snapshots were created years apart, and many are unrelated to each other. My best guess is the original volume was at generation 27709, at the time it was made into a seed device, and then replicated into a sprout. The original file system is gone, and what I have now is this sprout, now at generation 33935. I wonder if a similar effect happens with btrfs replace or btrfs dev add/rem? The leaf containing the ROOT_ITEM for an example snapshot from 2016 generation 27709 root_dirid 256 bytenr 1471499272192 level 2 efs 1 lastsnap 27697 byte_limit 0 bytes_used 92422144 flags 0x1(RDOLY) uuid d5d492f6-7421-c74b-a383-287caaac9a6c parent_uuid 305449bb-3580-614e-8f7e-e30c8385e0e6 ctransid 1004 otransid 1007 stransid 0 rtransid 0 That's a bit confusing. I guess otransid=ogen and ctransid is something else. And transid=generation. And also something else called lastsnap. All different values. The fs_tree show the transid for this snapshot is 27709, which matches the generation found in the ROOT_ITEM. -- Chris Murphy
