On Mon, Nov 27, 2000 at 02:24:03PM -0700, Hans Reiser wrote:
> using reiserfs over raid5 with 5 disks. This is unnecessarily suboptimal, it should
>be that parity
> writes are 20% of the disk bandwidth. Comments?
It would be great to have a number describing ``unnecessarily suboptimal''
more accurately ;)
Last I looked, the RAID-5 code would, upon performing a write:
1) Read a block from all disks but the parity and the destination disk
2) Calculate parity from those reads along with the new block
3) Write the new block and the new parity
However, I cannot see a reason why (1) cannot be replaced with reading the
old parity, and the old data block, then changing (2) to calculating parity
based on that information instead.
If we are reading the old block anyway (which I guess is often the case), then
we will make two reads and two writes, instead of N-1 reads and two writes.
But maybe the code changed (I don't think so), or maybe there are good reasons
for not doing this (?)
>
> Is there a known reason why reiserfs over raid5 is way worse than ext2. Does ext2
>optimize for
> raid5 in some way?
>
I don't know about ext2. It does have the ``stride'' option in mke2fs, but
that's about it. I don't remember either exactly what it does, or how
much of an impact it had on performance...
But the RAID layer does not optimize for ext2, that's for sure.
--
................................................................
: [EMAIL PROTECTED] : And I see the elder races, :
:.........................: putrid forms of man :
: Jakob �stergaard : See him rise and claim the earth, :
: OZ9ABN : his downfall is at hand. :
:.........................:............{Konkhra}...............:
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to [EMAIL PROTECTED]
