I read that that should be case but did not see such in practice. I created one volume without the recsize setting and one with. Than copied the same data to both (lots of small files). The 'du' report on the one without the recsize was significantly bigger than the one where I made it and in fact filled up before the other volume.
On Mon, 18 Aug 2008, Robert Milkowski wrote: > Hello Paul, > > Thursday, August 14, 2008, 9:25:45 PM, you wrote: > > PR> This problem is becoming a real pain to us again and I was wondering > PR> if there has been in the past few month any known fix or workaround. > > PR> I normally create zfs fs's like this: > > PR> zfs create -o quota=131G -o reserv=131G -o recsize=8K zpool1/newvol > > PR> and then just nfs export through /etc/dfs/dfstab. We deal with lots > PR> of small image files in our MRI data which is the reason for small > PR> recsize. > > If you are creating lots of small files and you are not worried about > accessing only small parts of them with a smaller block, then it > doesn't really make sense to reduce record size. With standard 128K > recsize if you create a small file, lets say 8K in size, the block > size used by zfs will be actually 8K. > > Unless you are doing this to work-around space maps issue. > > > -- --------------------------------------------------------------- Paul Raines email: raines at nmr.mgh.harvard.edu MGH/MIT/HMS Athinoula A. Martinos Center for Biomedical Imaging 149 (2301) 13th Street Charlestown, MA 02129 USA _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
