[NILFS users] Using nilfs to increase performance of NAND/SSD flash
burton at tailrank.com
Sat Feb 2 05:31:47 JST 2008
I'm playing with nilfs2 on an Mtron SSD 32G NAND Flash SATA drive.
NAND is GREAT for random reads but horrible for random writes.
A log filesystem fixes this problem because when the disk does random writes
will just read the block from the previous erase block and wrote a new block
with the data to the end of the drive, thereby simulating sequential writes.
I tested this theory by comparing nilfs2 to XFS on the mtron with a
random write to test. Sure enough nilfs2 beat out XFS by 8.5x and wrote at
about 50MB/s which is amazingly impressive.
The problem is that with I used the sysbench OLTP benchmark it was only
10% faster than an HDD running XFS.
Not too impressive.
The OLTP benchmark simulates a MySQL box doing lots of random DELETE,
and INSERT commands so you end up seeing lots or random writes to the disk.
In theory nilfs2 should be able to boost hte performance here but I didn't
this in practice.
I think what might be happening is that the consistent snapshot support
slowing down performance by keeping older blogs on disk and doing too many
sequential writes thereby seeing slower performance.
Is it possible to disable checkpointing to test my theory? I'm just about
dive into the source to test this out but I wanted to post here first to get
Location: San Francisco, CA
Work: http://spinn3r.com and http://tailrank.com
Fax: 1-415-358-419 PIN: 0092
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the users