Gluster users - Jul 2017 - RECOMMENDED CONFIGURATIONS - DISPERSED VOLUME

Hi

I'm looking for an advise to configure a dispersed volume.
I have 12 servers and would like to use 10:2 ratio.

Yet RH recommends 8:3 or 8:4 in this case:

https://access.redhat.com/documentation/en-US/Red_Hat_Storage/3.1/html/Administration_Guide/chap-Recommended-Configuration_Dispersed.html

My goal is to create 2PT volume, and going with 10:2 vs 8:3/4 saves a few
bricks. With 10:2 I'll use 312 8TB bricks and with 8:3 it's 396 8TB
bricks
(36 8:3 slices to evenly distribute between all servers/bricks)

As I see it 8:3/4 vs 10:2 gives more data redundancy (3 servers vs 2
servers can be offline), but is critical with 12 nodes? Nodes are new and
under warranty, it's unlikely I will lose 3 servers  at the same time (10:2
goes offline). Or should I follow RH recommended configuration and use
8:3/4?

Thank you.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.gluster.org/pipermail/gluster-users/attachments/20170731/4667c692/attachment.html>

Dmitri the recommendation from redhat is likely because it is recommended
to have the data stripes be a power of two otherwise there is a performance
penalty.

On 31 July 2017 at 14:28, Dmitri Chebotarov <4dimach at gmail.com> wrote:
> Hi
>
> I'm looking for an advise to configure a dispersed volume.
> I have 12 servers and would like to use 10:2 ratio.
>
> Yet RH recommends 8:3 or 8:4 in this case:
>
> https://access.redhat.com/documentation/en-US/Red_Hat_Storage/3.1/html/
> Administration_Guide/chap-Recommended-Configuration_Dispersed.html
>
> My goal is to create 2PT volume, and going with 10:2 vs 8:3/4 saves a few
> bricks. With 10:2 I'll use 312 8TB bricks and with 8:3 it's 396 8TB
bricks
> (36 8:3 slices to evenly distribute between all servers/bricks)
>
> As I see it 8:3/4 vs 10:2 gives more data redundancy (3 servers vs 2
> servers can be offline), but is critical with 12 nodes? Nodes are new and
> under warranty, it's unlikely I will lose 3 servers  at the same time
(10:2
> goes offline). Or should I follow RH recommended configuration and use
> 8:3/4?
>
> Thank you.
>
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://lists.gluster.org/mailman/listinfo/gluster-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.gluster.org/pipermail/gluster-users/attachments/20170731/bf77ee4b/attachment.html>