Ask Your Question

Best way to create HA SSD storage backend for Openstack?

asked 2017-05-31 03:27:18 -0500

OsQ gravatar image


we have an Openstack cloud and beside them 2 storage nodes with 4TB of SSD disks on each. The purpose would be to produce high performance volumes for Openstack instances. Only 1 node will serve as storage, other node would be HA for that. Any recommandations or solutions? I have now looked in to GlusterFS and Quadstor, but they don't seem to be valid solutions.

edit retag flag offensive close merge delete

2 answers

Sort by ยป oldest newest most voted

answered 2017-05-31 19:56:32 -0500

A common solution is Ceph. Since it is object storage at its core, it is inherently HA. You would normally want three storage servers and implement three-way replication, but two should be OK. Or you check out erasure coding, also supported in Ceph.

edit flag offensive delete link more

answered 2017-06-07 05:25:54 -0500

OsQ gravatar image

Thank you for your answer. CEPH was our thoughts also, but we started to think that does CEPH profit from SSD disks. Does it add any performance to the storage? If you compare CEPH with SAS disks versus CEPH with SSD disks.

edit flag offensive delete link more


The answer is "it depends". See

Bernd Bausch gravatar imageBernd Bausch ( 2017-06-07 16:39:56 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower


Asked: 2017-05-31 03:27:18 -0500

Seen: 187 times

Last updated: Jun 07 '17