We backup our ADVFS filesets by creating a clone fileset and
vdumping the clone. On filesets with heavy write activity during
a long-running backup, removing the clone fileset can take quite a
long time -- this morning one of ours took nearly an hour.
I understand that these cloned filesets inherently require
significant overhead during the lifetime of the clone, since
before-images of any modified blocks must be saved in the clone.
But I would have thought the rmfset operation on a clone would
be simply a matter of resetting some pointers and freeing the
before-image blocks. Obviously it's more complicated than that.
Could someone who understands the internals enlighten me?
Also, suppose I modified our backup script to do the rmfset in the
background and go on to backup the next fileset. Should I expect
that to shorten the overall backup time for multiple filesets, or
would the rmfset just compete for disk access with the next vdump
and its associated cloning overhead? Most of our filesets share
one large file domain, and we're heavily CPU-overloaded during
these backups. This is happening on DU 4.0A.
Thanks!
--
-- Phil Rand <prand_at_spu.edu>, aka <postmaster_at_spu.edu>
-- http://www.spu.edu/users/prand (206) 281-2428
-- Computer and Information Systems
-- Seattle Pacific University
-- 3307 3rd Ave. W., Seattle, WA 98119
----------------------------------------------------------------------
"One person CAN change the world, but most of the time,
you probably shouldn't." -- Marge Simpson
Received on Thu Feb 25 1999 - 21:52:04 NZDT