On Saturday July 30, scott-raid@riskboys.com wrote:
I'm doing a big rsync to my 5 disk RAID 5 partition and am seeing an
uneven loading of my disks. The files are almost all huge video files.
Has anyone got any idea what I could have done wrong?
md6 : active raid5 sdi1[4] sdh1[3] sdg1[2] sdf1[1] sde1[0]
976783360 blocks level 5, 128k chunk, algorithm 2 [5/5] [UUUUU]
The filesystem is ext3 created with
mkfs.ext3 -R stride=32 /dev/md6
I suggest you ask on ext3-users@redhat.com.
I have seen references to this sort of thing before. I think some
metadata gets aligned despite the '-R stride=' flag.
Also, I think the stride should be 128 (32 fs-blocks per chunk, 4
chunks per stripe).
NeilBrown
On Sun, 2005-07-31 at 00:33, Neil Brown wrote:
On Saturday July 30, scott-raid@riskboys.com wrote:
quoted
I'm doing a big rsync to my 5 disk RAID 5 partition and am seeing an
uneven loading of my disks. The files are almost all huge video files.
Has anyone got any idea what I could have done wrong?
md6 : active raid5 sdi1[4] sdh1[3] sdg1[2] sdf1[1] sde1[0]
976783360 blocks level 5, 128k chunk, algorithm 2 [5/5] [UUUUU]
The filesystem is ext3 created with
mkfs.ext3 -R stride=32 /dev/md6
I suggest you ask on ext3-users@redhat.com.
I have seen references to this sort of thing before. I think some
metadata gets aligned despite the '-R stride=' flag.
Also, I think the stride should be 128 (32 fs-blocks per chunk, 4
chunks per stripe).
NeilBrown
Speaking of which, how does one determine the stride size? If I do:
mdadm -C /dev/md0 -l 5 -c 64 -n 2 /dev/sd{ab}1
mke2fs -j -b 4096 -R stride=??? /dev/md0
What should the stride size be--64x4? Is the default chunk size still
64K?
Thanks,
Forrest