In the current design of zfs(1m), nearly all of the operations are limited to working on a single filesystem at a time. This places some limitations on the usability of zfs(1m) when using it with scripts. For example, you cannot do:
# zfs list -H -o origin | egrep -v '^\-$' | xargs zfs destroy Instead, you need to do this: # zfs list -H -o origin | egrep -v '^\-$' | sh -c 'while read i; do zfs destroy "$i"; done' (because not everyone uses sh as their shell...) As an experiment, I modified zfs_main.c to allow destroy to work in a manner that is conducive to being used with xargs - patches attached. It's a relatively simple change but there are important questions, such as does it fail on the first error or continue on, etc. But I'm sure there was a reason why it was designed and implemented the way it is rather than with what my patch suggests? Darren -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: zfs_main.c.patch URL: <http://mail.opensolaris.org/pipermail/zfs-code/attachments/20080908/30bdf0e2/attachment.ksh>
