>> ... >> Pieter Bowman <bow...@math.utah.edu> wrote: >> >> > The problem seems to be caused by the changing of the inode of the >> > root of that filesystem. The inode for the test filesystem's root >> > directory is 3, the inode for various snapshots are numbers like: >> > >> > 281474976666177 >> > 281474976671479 >> > 281474976673971 >> >> So it seems that there is a bug in your specific ZFS snapshot implementation. >> >> The original ZFS snapshot implementation behaves as expected: >> ...
Solaris 10 and 11 (which we still use) have the same inode number for snapshots and filesystem. I just checked on a FreeBSD 12 system, which has the same inode numbers. Apparently the ZFS on Linux has chosen not to do this. I found one comment about this by behlendorf: https://github.com/zfsonlinux/zfs/issues/4968 @svirusxxx thanks for digging in to this and filing an issue. Based on the debugging you've posted the issue appears to be that tar detects that the root inode number of mounted snapshot has changed. This was done intentionally by ZFS and that root inode number is used to encode exactly what snapshot needs to be auto-mounted there. If tar can be updated to ignore just that inode number for ZFS that may be the best solution. Thanks for the plug for star. I may look at switching to star using the amanda-applications amstar interface. Pieter
