Ask Your Question

ephemeral disk nfs shared storage. What about HA?

asked 2014-10-01 02:10:23 -0500

Marco Marino gravatar image

Hi, here -> there is an example of live migration with shared (nfs) storage. It works well, i've tested it on my 10 compute nodes private cloud. But what happens if the storage server (with glusterfs) goes down? All ephemeral disks of all VMs will be corrupted, even if we use an HA solution beacuse the "switch-time" of services from storage1 (that goes down) to storage2 causes a data corruption of the ephemeral disks. Is this right? How can i solve this? Should i use some kind of active/active HA configuration for glusterfs or data corruption is unavoidable? Thanks

edit retag flag offensive close merge delete


Use Ceph? I would recommend using Ceph if you can.

mpetason gravatar imagempetason ( 2014-10-01 10:49:53 -0500 )edit

2 answers

Sort by ยป oldest newest most voted

answered 2015-09-11 07:12:21 -0500

visbits gravatar image

He's talking about his libvirt config files. I think its silly that openstack isn't storing the configuration data in the database. I'm thinking about using CephFS to store my libvirt config files, all of my vm storage is ceph/iscsi.

edit flag offensive delete link more

answered 2014-10-01 03:04:01 -0500

nd gravatar image

If you create a cluster of nodes for your glusterfs then it will provide resilient storage and failure on one glusterfs storage node should not cause unavailability of your volumes. You can refer to this link [and surely there would be many others], which shows "How To Create a Redundant Storage Pool Using GlusterFS on Ubuntu Servers" (


edit flag offensive delete link more

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower


Asked: 2014-10-01 02:10:23 -0500

Seen: 539 times

Last updated: Sep 11 '15