The real problem for us is down to the fact that with ufsdump and ufsrestore
they handled tape spanning and zfs send does not.
we looked into having a wrapper to "zfs send" to a file and running gtar (which
does support tape spanning), or cpio ... then we looked at the amount we
started storing
I wrote a simple script to dump zfs snapshots in a similar way to
what I did with ufsdump. In particular, it supports dump levels.
This is still very crude, but I'd appreciate comments. Does what
I'm doing make sense, or do I misunderstand zfs?
Thanks.
-
#!/bin/ksh
#
# XXX - real option pa
