I benchmark a lot on various hardware and software configurations. When I
started 2 years back I went along with the general consensus that XFS is
faster than ext4 for MySQL. I recently had the opportunity to see how much
of a difference, if any, it made. I didn't find much, especially on SSD.
The benefit of ext4 on SSD (on newer kernels) is that it supports the TRIM
functionality (add discard to your /etc/fstab file for the particular file
system). I don't believe TRIM is supported for XFS.
On Wed, May 22, 2013 at 2:55 AM, Rafał Radecki <radecki.rafal@stripped>wrote:
> Hi All.
> I use mysql/perconna/maria on my production CentOS 6 Linux servers. I
> currently try to choose the default filesystem for partitions with mysql
> data. Some time ago (previous dba) reiserfs was the choice but now it is
> not in the kernel and the main author is in prison.
> From what I've read xfs and ext4 are valid choices and performance
> benchmarks over the web show that they are "comparable" (no clear winner).
> I've also read that with every new kernel there can be changes in
> performance in every filesystem ( for example
> From your experiences: which filesystem to choose for a mysql db? Is ext4
> or xfs better? Or is it more a case of proper filesystem tuning to my
> workload? Any articles worth reading which you can recommend?
> Best regards,