[email protected] (Volker A. Brandt) writes: > Hi Harry! > > >> > Both will work. Any dataset snapshot can be sent/received separately, >> > if you want. But you can also use a recursive send where the dataset >> > itself and all "child datasets" will be included in the stream. Check >> > out the "-r" and "-R" flags in the zfs(1M) man page. >> >> I could not find an -r (lowercase) for send sub-command (maybe you are >> only refering to zfs main cmd). > > No, it was a Solaris-ism that doesn't exist in illumos. Sorry for not > having checked it before sending the mail. > >> But experimenting with send -R I seem to have >> gotten myself into some trouble. >> >> I attempted to: >> zfs send -R rpool/export/home/MYUSER |ssh oi0 zfs recv p12x > > OK, so you are transferring an entire dataset, not a snapshot. Not > sure how this could have worked, as the -R flag only applies to > snapshots, not datasets.
More sloppy work on my part, I actually did send a snapshot just shorted the info in my post. [...] Snipped the result of my misleading information But the good thing is that even after being totally mislead by me you still managed to solve my issue. > > zfs send -R rpool/export/home/MYUSER@now | > ssh oi0 zfs recv -e -o mountpoint=/export/home/MYUSER p12x/export/home > > You could leave out -R since there is nothing to replicate if you only > have one dataset with one snapshot. > > The "-e" will tell zfs recv to just use the MYUSER part of the sent > dataset name. Note that p12x/export/home must already exist. > > If you want to be sure everything is copied before mounting it, use > "zfs recv -u ..." so that datasets aren't automatically mounted on the > receiving side. > > > Hope this helps -- Volker Very much so. And thanks for being persistent even in the face of my sloppy, mis-managed plea for help. _______________________________________________ openindiana-discuss mailing list [email protected] https://openindiana.org/mailman/listinfo/openindiana-discuss
