[Linux-PowerEdge] KVM Disk Format Options
danno at umich.edu
Wed Feb 20 14:53:31 CST 2013
The answer may be different depending on your storage vendor; some of
them have better NFS performance (i'm looking at you, netapp) and others
are better at iscsi. So, I'd ask EMC what they recommend.
Jose, is it really true that you need iscsi to do live VM migration with
KVM? That is not true with vmware; NFS works too.
Jose_De_La_Rosa at Dell.com wrote:
> I don't know of existing Dell whitepapers that have studied
> performance of block vs. file-based storage, but the statement below
> makes sense, as using block storage avoids having to use the file
> system layer on the hypervisor. When using iSCSI-based storage pools
> and volumes (required for live VM migration) the default storage type
> will be block based.
> *Jose De la Rosa*
> *Linux Engineering***
> *Dell****| Enterprise Solutions Group***
> *From:*linux-poweredge-bounces-Lists *On Behalf Of *Mesh
> *Sent:* Wednesday, February 20, 2013 1:04 AM
> *To:* linux-poweredge-Lists
> *Subject:* [Linux-PowerEdge] KVM Disk Format Options
> Good day,
> We are implementing a multi-node system based on KVM that will have
> high disk I/O and I'm doing some research so that I can
> advise whether to use block devices (LUNs) for our VMs, or image
> files. I'm getting confusing messages, depending on where I look,
> including some Dell white papers that I could access.
ICPSR Computing & Network Services
University of Michigan
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Linux-PowerEdge