This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
Home Products Support Community News


RES: RES: [Xen-users] Shared volume: Software-ISCSI or GFS or OCFS2?

Sorry my bad.

I don't have 03 storages. I have 01 storage and 03 hosts accessing them.

My goal is (regardless technology): Provide minumum downtime with Xen. I
mean, IF a host has a harwdare malfunction, I can Just start the domU in
another host without do anything else unless mount/Access the path.



-----Mensagem original-----
De: Rustedt, Florian [mailto:Florian.Rustedt@xxxxxxxxxxx] 
Enviada em: quinta-feira, 20 de novembro de 2008 10:34
Para: Bruno Bertechini
Cc: xen-users@xxxxxxxxxxxxxxxxxxx
Assunto: AW: RES: [Xen-users] Shared volume: Software-ISCSI or GFS or OCFS2?

 ...my experience with LVM-snapshoting a live-system is, that it freezes
with XFS-filesystem.

While the snapshot is running, the disk is "frozen" for some moments and XFS
is "intelligent" enough to recognize this and throws an error with the
result, that the filesystem is shut down. This is normally a good behaviour
to prevent a system to be corrupted when the disk fails. In this case it
prevents snapshots from live-systems ;)

This is known on the mailling-lists, btu no near solution in sight.

I don't know, if other filesystems are better for this - better means more
stupid ;) - but i wouldn't rely on it.

In the moment I am thinking of using
DRBD -> LVM -> (OCFS2 mounted directly) or (XFS distributed via NFS) for our
solution, because i want to mount some disks multiple times.

As far as i understood now(please correct me), there are two principles to
do this:

1. file based, distributing file-based disks with some technology
2. blockdevice based, distributing blockdevices with some technology

In any case, you must decide how you want the three storages to be combined
before you can start.
In case 2, your result could be a single blockdevice on which you could use
lvm for easy partitioning.
In that case, you could mount the lv's directly with any filesystem on them,
as long as you don't want to share or snapshot them. Perhaps the commercial
DRBD+ (linbit.com) is interesting for this with three storages.

Regards, Florian
-----Ursprüngliche Nachricht-----
Von: Bruno Bertechini [mailto:bruno.bertechini@xxxxxxxxxxx] 
Gesendet: Donnerstag, 20. November 2008 12:57
An: 'John Haxby'
Cc: 'Bastian Blank'; Rustedt, Florian; xen-users@xxxxxxxxxxxxxxxxxxx
Betreff: RES: RES: [Xen-users] Shared volume: Software-ISCSI or GFS or

Well.. Let me use this thread and popup my environment and ask for

We have a EMC storage and 03 hosts (all with fibre channel adapters).

Choice 1 :

Regarding Xen backend : Files or LVM ? I'm thinking to use LVM to use online
resize/snapshots. Are there huge performance differences?

I can't define a infrastructure for EMC/NFS/OCFS/etc without choose before
select the appropriate backend.

What do you suggest guys?



-----Mensagem original-----
De: John Haxby [mailto:john.haxby@xxxxxxxxxx] Enviada em: quarta-feira, 19
de novembro de 2008 11:30
Para: Bruno Bertechini
Cc: 'Bastian Blank'; 'Rustedt, Florian'; xen-users@xxxxxxxxxxxxxxxxxxx
Assunto: Re: RES: [Xen-users] Shared volume: Software-ISCSI or GFS or OCFS2?

Bruno Bertechini wrote:
> Can someone explain why we should use NFS? It is fairly slow and unsecure.
It depends.  If you're using a NetApp server or an EMC Celerra then NFS 
isn't slow and it's secure "enough".   The big advantage of NFS is that 
it is easy to get working and it's well understood by a lot of people.

> Why not use some clustered FS ? 
NFS is a clustered file system -- it's a file system that's visible 
across a cluster.   If you mean an HA cluster, then yes, you can use 
something like ocfs2 or gfs/gfs2.   These things, however, are designed 
when you want concurrent access to individual files within the file system
and for Xen disk image files you don't normally want that -- unless you're
using a clustered file system in the guests of course.

The right file system for you depends, in detail, what you want to do.

> Are there alternatives ?

For Xen virtual disk files there are several alternatives: iSCSI and nbd 
both provide access to logical devices across a standard network.   That 
EMC Celerra I mentioned supports iSCSI as well as NFS and, depending on 
what you're doing, you may prefer that.   If you have a fibrechannel SAN 
or even infiniband you can use those as well.

For my non-production purposes I use NFS because it's easy and fast (that
is, the quickest way to get data between two machines over the 
gigabit LAN is using NFS).   If I were setting up a production cluster 
then I would start with some shared non-local storage and I'd probably be
looking at NFS again or iSCSI and the choice then depends on what the 
disk array box supports and is good at.   For a system built entirely 
out of stock PCs, well, I wouldn't. I wouldn't build a production cluster of
any size that way: I want proper storage.

What you do depends on what you're planning to use these machines for and
how much time and money you're spending.

IMPORTANT: The contents of this email and any attachments are confidential.
They are intended for the 
named recipient(s) only.
If you have received this email in error, please notify the system manager
or the sender immediately and do 
not disclose the contents to anyone or make copies thereof.
*** eSafe scanned this email for viruses, vandals, and malicious content.

Xen-users mailing list