FYI - Here are the fixes available in builds 32 and 33. Most notable of these is the improvement in CLI performance, with more on the way in build 34 (6377670 being the most noticeable). For those of you with lots of filesystems, you should see an improvement. A number of performance fixes (DNLC caching, 16K ZAP leaf block size, dbuf caching in the ARC) are due in builds 34-35, which should increase performance across the board, including CLI performance. - Eric 6343802 zpool(1M) should make better use of libdiskmgt 6347801 internal error: unexpected error 79 at line 1617 of ../common/libzfs_dataset.c 6350233 Cannot rename filesystem 6354197 setting env variable NOINUSE_CHECK causes zpool to core dump 6354804 The file''s ACL was changed when cp it from one ZFS file system to another one. 6355681 cannot remove xattr even while write_attr is granted 6362672 import gets confused about overlapping slices 6362908 nfsv4-test: RENAME does not check PERM with NFSv4/ZFS 6364126 cannot rollback when a child filesystem is created after a snapshot is taken 6364129 need to unmount any child datasets before doing the rollback 6364275 internal error: disk configuration error 19 at line 518 of file zpool_vdev.c 6364582 need to fixup paths if they''ve changed 6365038 Legacy umount [dir] will also unmount filesystems mounted in subdirectories of [dir] 6366258 SUNWzfsr breaks in whole root zones 6366467 Different x_attr behavior while ACL against the same user by ''user:'' or ''owner@'' 6367517 32-bit kernel VA exhaustion with logging ufs atop zvol 6368111 ZFS returns EACCES in preference to EEXIST 6368181 some memleaks found related to zfs mount 6372094 zil_commit() may be called recursively and deadlock with itself 6372107 legacy "umount" command unmount filesystems in the subdirs when zfs mountpoint is a regular path 6372549 nfsv4 server returns OK for SETATTR of size on a symlink over ZFS, INVAL is expected 6377671 zfs mount -a shouldn''t bother checking snapshots 6377673 ''zfs mount -a'' should discover the proper mount order 6378361 ''zfs share -a'' needs to avoid expensive checks during boot 6378377 zfs_get_stats() is too expensive 6378388 zfs_for_each() iterates unnecessarily -- Eric Schrock, Solaris Kernel Development http://blogs.sun.com/eschrock