Hi,
you can convert a normal volume into tiered volumes in Gluster. once a
volume (for example volume1) is converted into tiered volumes it will
have two tiers (Hot and Cold).
So naming the volume hot tier, cold tier doesn't make much sense.
About specifying the tiers. The volumes you currently have will be the
cold tier. so if you want the cold tier to be replicate, then create a
replica volume and then attach your
Hot tier (of the volume type you need) to it.
About suggestion for tiered volume, The cold is usually a EC volume
and the hot is a replica volume.
On Sat, Feb 10, 2018 at 12:23 AM, David Brown <dbccemtp at gmail.com>
wrote:> Hello everyone.
> I have a new GlusterFS setup with 3 servers and 2 volumes. The
"HotTier"
> volume uses Nvme and the "ColdTier" volume uses HDD's. How do
I specify the
> tiers for each volume?
>
> I will be adding 2 more HDDs to each server. I would then like to change
> from a Replicate to Distributed-Replicated. Not sure if that makes a
> difference in the tiering setup.
>
> [root at Glus1 ~]# gluster volume info
>
> Volume Name: ColdTier
> Type: Replicate
> Volume ID: 1647487b-c05a-4cf7-81a7-08102ae348b6
> Status: Started
> Snapshot Count: 0
> Number of Bricks: 1 x 3 = 3
> Transport-type: tcp
> Bricks:
> Brick1: Glus1:/data/glusterfs/ColdTier/brick1
> Brick2: Glus2:/data/glusterfs/ColdTier/brick2
> Brick3: Glus3:/data/glusterfs/ColdTier/brick3
> Options Reconfigured:
> transport.address-family: inet
> nfs.disable: on
> performance.client-io-threads: off
>
> Volume Name: HotTier
> Type: Replicate
> Volume ID: 6294035d-a199-4574-be11-d48ab7c4b33c
> Status: Started
> Snapshot Count: 0
> Number of Bricks: 1 x 3 = 3
> Transport-type: tcp
> Bricks:
> Brick1: Glus1:/data/glusterfs/HotTier/brick1
> Brick2: Glus2:/data/glusterfs/HotTier/brick2
> Brick3: Glus3:/data/glusterfs/HotTier/brick3
> Options Reconfigured:
> auth.allow: 10.0.1.*
> transport.address-family: inet
> nfs.disable: on
> performance.client-io-threads: off
> [root at Glus1 ~]#
>
>
> Thank you all for your help!
>
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://lists.gluster.org/mailman/listinfo/gluster-users
--
Regards,
Hari Gowtham.