Hi Volker,
Hello All,
I am new to this list.
I am planning do create a cluster with several Gluster nodes. I get to
the following questions:
a) Is it possible to build a cluster with some of the bricks as SSD and
some of them with HDD? In the end I want two separate pools one fast
pool with SSD storage and a second slow pool with HDD. Is there a way to
define a scenario like this in Gluster? The question is not related to
SSD caching.
Cluster in GlusterFS is only required to create group of nodes which can be used to provision Volumes as required. Disks are not tied to Cluster. Volumes can be created as required. For example, Create a Volume using HDD and create another volume using SSD. (I am not sure this flexibility is available in Heketi or not, may be that supports tagging the devices and provision using those tags)
b) Or, shall I spilt this in two clusters of Gluster, one fast the
another one slow.
As answered above, Splitting the Cluster is not required. Create two Volumes.
My imagination is to build something like this:
Kubernetes Cluster 1 ... Kubernetes Cluster N
| |
storage class slow <-> storage class fast
-----------------------------------------------------
\ /
Heketi
/ \
Gluster 1 (fast) Gluster 2 (slow)
https://kadalu.io can be used as alternative to Heketi to use GlusterFS with Kubernetes. Create two Storage pools one with SSD and another with HDD, and create Storage class to use these pools.
|
________
Community Meeting Calendar:
Schedule -
Every 2nd and 4th Tuesday at 14:30 IST / 09:00 UTC
Bridge: https://bluejeans.com/441850968
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
https://lists.gluster.org/mailman/listinfo/gluster-users
