Gluster users - Nov 2016 - Improving IOPS

Il 05/11/2016 11:45, Lindsay Mathieson ha scritto:
> Six of one, half a dozen of the other. With RAIDZ10 I can lose one 
> disk, maybe two without losing the brick. Replacing the disk is very 
> easy and fast.
>
>
> Resilvering one disk in a raid10 set is a lot quicker and less impact 
> than resyncing a brick via the network.
Ok, I wasn't clear enough.
Do you have a single RAIDZ10 or multiple RAIDZ1 ?
In a single RAIDZ10, if you totally loose a mirror (thus, both disks 
from the same RAIDZ1 set), you loose the whole RAID10.

As you are using Gluster, is much better to use multiple RAIDZ1, each 
set as a single bricks. Gluster will aggregate space, so
you'll get the same space as a single RAIDZ10, but in case of a mirror 
failure, you have to recover only that single mirror.

On 5/11/2016 9:02 PM, Gandalf Corvotempesta wrote:
> Ok, I wasn't clear enough.
> Do you have a single RAIDZ10 or multiple RAIDZ1 ?
Single RAIODZ10, one brick per node
> In a single RAIDZ10, if you totally loose a mirror (thus, both disks 
> from the same RAIDZ1 set), you loose the whole RAID10. 
Yah, I get that. For me willing to risk loosing the entire gluster node 
and having to resync it, I see the odds as pretty low vs just losing one 
disk in the RAID10 set and resilvering it locally.


-- 
Lindsay Mathieson