Red Hat Enterprise Virtualization now supports Red Hat Storage, a new distributed file system based on the Gluster...
file system, but its use cases are limited.
Put simply, Red Hat Storage allows administrators to create one centrally managed storage pool, where the actual data is stored on different physical locations. It's an interesting technology, but in many cases it just doesn't make sense to use Red Hat Storage and you're better off using a classical storage solution like iSCSI or Fibre Channel.
Storing VM files in RHEV
In a typical RHEV environment, virtual machine (VM) files are stored in one central location, typically an iSCSI or Fibre Channel storage area network (SAN). If, for some reason, your servers can't access that storage location, they will no longer be able to run VMs. Red Hat Storage could help you avoid this problem by replicating the storage volume.
Apart from replicated volumes, Red Hat Storage also has an option that creates distributed and striped volumes, or any combination of the three volume types (distributed, replicated and striped volumes). Distributed volumes may not be useful to most data centers, because with this volume type you never know which physical server hosts your files, creating a troubleshooting nightmare.
In this article, I'll focus on replicated volumes, but the same constraints apply to other volume types.
There are two main use cases for Red Hat Storage in a RHEV data center. One is for large environments where Red Hat Storage can be used as an alternative to SAN replication. The other helps you set up a small RHEV site without using an external SAN. Let's look at the advantages and limitations in both cases.
Using Red Hat Storage for disaster recovery
The most obvious reason to use Red Hat Storage in RHEV is that it automatically replicates VM image files between different sites. That allows you to set up an affordable disaster recovery scenario where, if one site goes down, the VM files are still available at the other site. This is a great feature for some larger businesses but is probably not the best solution for every environment.
Some SAN filers have built-in replication features. If you're looking for file replication between sites and your SAN offers that functionality, there is no need for Red Hat Storage. If, however, your SAN does not offer this functionality, or the licensing cost for SAN mirroring is too high, Red Hat Storage might be an option.
Red Had Storage as a SAN alternative
Another potential use case for Red Hat Storage could be as an alternative to a SAN at a small site. The reliability of this approach depends entirely on your hardware setup: It will only work if your server, storage and networking hardware is already redundant. In this case, you could consider creating a distributed, replicated Red Hat Storage volume. To create this setup, you need a storage brick on each server in the RHEV cluster and to connect each server to this RHEV volume. Data will be replicated between nodes, so if one node fails, the other nodes will be capable of accessing all VM files.