> Sorry if duplicated...
> Thanks guys, let me expose my problem. I'm prototyping a fault tolerance
> server (byzantine fault tolerance). Vms need do comunicate each order for
> make decisions about requests. My host will be out of network (cant access
> anything) and my Vms can comunicate to the external world, but the decision
> need to be local. Yes, I could comunicate via network inside VM to exchange
> information but my propose is to have a reliable local share to do it. In
> another work a friend used a Virtualbox folder share to do it, but now I
> need increase performance and use Xen to do it.
> About clustered filesystems, any options to use it in a normal block device
> like Sata HD or SCSI? I understand..."clustered"... but any option?
> Thanks and I will get more information about clustered filesystems
> 2011/2/14 Javier Guerra Giraldez <javier@xxxxxxxxxxx
>> On Mon, Feb 14, 2011 at 1:37 PM, lucianobarreto@xxxxxxxxx
>> > I need to share some files between VMs. This files will be used to
>> > transfer
>> > some information (read/write). But I need do it without any network
>> > resource
>> > (NFS or others). I've tried to do it sharing a partition just for test
>> > proposes, but i see that when i create a file on one VM another can't
>> > see it
>> > and there isnt any concurrence in this approach.
>> > Anyone can help me??
>> to share files, you need a shared filesystem. there are two main
>> classes of these:
>> - network filesystems: NFS, Samba, 9p, etc. these work really well;
>> you shouldn't reject them without good reason.
>> - clustered filesystems: GFS, OCFS2, CXFS, etc. they're designed for
>> SAN systems where several hosts access the same storage box. in VM
>> case, if you create a single partition accessible from several VMs you
>> get exactly the same situation, (shared block device) and need the
>> same solution.
>> what definitely won't work is to use a 'normal' filesystem (ext3/4,
>> XFS, ReiserFS, FAT, HPFS, NTFS, etc) on a shared partition (just like
>> it won't work in a shared block device). Since every filesystem
>> aggressively caches metadata to avoid rereading the disk for every
>> access, a VM won't be 'notified' if another one modifies a directory,
>> so it won't 'notice' any change. and worse, since now the cached
>> metadata isn't consistent with the content of the disk, any write will
>> result in a heavily corrupted filesystem.
>> better go with NFS
>> Xen-users mailing list
> Luciano Barreto
> Xen-users mailing list