[libvirt-users] using shared storage with libvirt/KVM?

[David Ehle]

Hello,

I've spent a few days googling and reading documentation, but I'm looking 
for clarification and advise on setting up KVM/libvirt with shared 
storage.

I have 2 (for now) Ubuntu Karmic systems with KVM/virsh/virt-manager set 
up and running.

I have a storage server that can do NFS/iSCSI/samba/ etc.

I am trying to figure out the best way to set things up so that I can run 
several (eventually) production linux guests (mostly debian stable) on the 
2 KVM host systems and be able to migrate them when maintenance is 
required on their hosts.

Performance is of medium importance, stability/availibilty is pretty high 
priority.

A common opinion seems to be that using LVs to hold disk images gives 
possibly the best IO performance, followed by raw, then qcow2.

Would you agree or disagree with this? What evidence can you provide?


I am also rather confused about using shared storage, and what options can 
be combined with this.

I have succesfully made an iscsi device available to 
libvirt/virsh/virt-manager via an XML file + pool-define + pool-start. 
However the documentation states that while you can create pools through 
libvirt, volumes have to be pre-allocated:
http://libvirt.org/storage.html
"Volumes must be pre-allocated on the iSCSI server, and cannot be created 
via the libvirt APIs."

I'm very unclear on what this means in general, and specifically how you 
preallocate the the Volumes.

When I make a iscsi pool available via something like this:
       <pool type="iscsi">
         <name>virtimages</name>
         <source>
           <host name="iscsi.example.com"/>
           <device path="demo-target"/>
         </source>
         <target>
           <path>/dev/disk/by-path</path>
         </target>
       </pool>

In virt-manger I can see it as a pool, but when I try to select it for 
where to create an image, it uses the pool AS A Volume (I think).

That brings me to my next question/misunderstanding...

If you are using shared storage (via NFS or iSCSI) does that also mean you 
MUST use a file based image rather than an LVM LV?  NFS makes directories 
available, not devices.  You can make an unformated "raw" LV available via 
iSCSI, but its not seen on the far side as a LV but instead as a scsi disk 
which you need to partition. You can make a PV>VG>LV out of that and then 
make your new LV available to libvirt/KVM but then your stacking a heck of 
a lot of LVM/LVs/partion tables up which seems pretty dubious.  I'm also 
not sure that the stack would be visable/available to the second system 
using that iSCSI device.  (I'm pretty new to using iSCSI as well.)

Redhat provides pretty good documentation on doing shared storage/live 
migration for NFS:
http://www.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/5.4/html/Virtualization_Guide/chap-Virtualization-KVM_live_migration.html
But unfortunately the section on Shared Storage with iSCSI is a bit 
lacking:
"9.1. Using iSCSI for storing guests
This section covers using iSCSI-based devices to store virtualized 
guests." and thats it.

For maybe 6-10 guests, should I simply just be using NFS? Is its 
performance that much worse than iSCSI for this task?

So since KVM/libvirt is getting used quite often in production now - but 
some simple things are not making any sense to me - I'm guessing that I 
have some really basic misunderstandings of how to do shared storage / 
Pools/ Volumes for migration.

Would one of the kind readers of this least be willing to correct me?

Please CC to my original email address as well as I am not yet a regular 
subscriber.

Thanks again!

David.