AoE barrier support

I remember following this thread with interest a couple years ago:

http://lkml.indiana.edu/hypermail/linux/kernel/0909.1/01208.html

but I never saw the barrier support turn up in a release. Did it get
implemented and I missed it somehow? I don't see it in the changelog or
anything relevant in the source. I recently heard from someone who had
filesystem corruption and they seemed to believe it would have been
avoided if the AoE initiator had barrier support. If the AoE initiator
they are running has it already (don't have his version number on hand
at the moment) then they are barking up the wrong tree. If write barrier
support is implemented I want to make sure my initiator has it also.

The code in the kernel patch in the above link does not seem to be
present in the current aoe6-76.tar.gz either but I could be
misunderstanding what is going on here. I thought if it was sent to the
kernel it would have been in the standalone kernel module distribution
already.

Thanks for any light you can shed on this for me!

--

-- 
Tracy Reed
http://tracyreed.org

------------------------------------------------------------------------------
Xperia(TM) PLAY
It's a major breakthrough. An authentic gaming

Bond Issues

We have our CoRaid devices setup on a Cisco 6500 switch with dual 6548A 
blades and standard bonding. I was reading that the AoE protocol is 
incompatible with several vendors bonding methods for the 802.3ad. What 
method are you using for Cisco switches that are working well?

Kris

------------------------------------------------------------------------------
Fulfilling the Lean Software Promise
Lean software platforms are now widely adopted and the benefits have been 
demonstrated beyond question. Learn why your peers are replacing JEE 
containers with lightweight application servers - and what you can gain 
from the move. http://p.sf.net/sfu/vmware-sfemails

Re: Bond Issues

On Sat, Apr 23, 2011 at 12:15:24AM -0500, NCServ Support spake thusly:
> I was reading that the AoE protocol is incompatible with several vendors
> bonding methods for the 802.3ad. What method are you using for Cisco switches
> that are working well?

Where did you read this?

I have used AoE with 802.3ad on both Cisco and HP switches and have had no
problems with it.

--

-- 
Tracy Reed

------------------------------------------------------------------------------
Fulfilling the Lean Software Promise
Lean software platforms are now widely adopted and the benefits have been 
demonstrated beyond question. Learn why your peers are replacing JEE 
containers with lightweight application servers - and what you can gain 
from the move. http://p.sf.net/sfu/vmware-sfemails

_______________________________________________
Aoetools-discuss mailing list
Aoetools-discuss@...
https://lists.sourceforge.net/lists/listinfo/aoetools-discuss

Re: Bond Issues

> -----Original Message-----
> From: Tracy Reed [mailto:treed@...]
> Sent: Saturday, April 23, 2011 1:57 AM
> 
> On Sat, Apr 23, 2011 at 12:15:24AM -0500, NCServ Support spake thusly:
> > I was reading that the AoE protocol is incompatible with several
> > vendors bonding methods for the 802.3ad. What method are you using
for
> > Cisco switches that are working well?
> 
> Where did you read this?
> 
> I have used AoE with 802.3ad on both Cisco and HP switches and have
had no
> problems with it.

The CORAID models I have used do not appear to support bonding on the
host<->switch links.  I don't think this is an intrinsic limitation of
the AoE protocol, but rather the network implementation on the CORAID
appliances.  That doesn't prevent you from using bonding between
switches, or from the switch to the initiator, but any lack of
redundancy on the target creates a point of failure for the storage
network as a whole.

But the real question is, why try to use bonding at all?  The multipath
implementation from CORAID works well, is supported, and (in our own
testing) seems superior in both performance and reliability to Ethernet
bonding.  It's also easier to setup.

-Jeff

Informations about ggaoed

Hello,

I am in the process of building a virtualisation platform based on Xen
and AoE.
Basicly I want 2 Xen hosts witch boot on an AoE SAN (root
device is /dev/etherd/ex.x). Additionnaly each virtual machine will have
its filesystem stored on the AoE SAN as well and exported to the dom0s.

I was more familiar with vblade until now, but i read here and there
that vblade could missbehave when serving a lot of devices.

So I now consider using ggaoed, which gives me good results for now.
Nevertheless I have few questions regarding ggaoed:

- Am I right to think that the design of ggaoed is more suited to
serve a lot of devices ( multithread V  forking  model)?

- Is ggaoed still in active developpement? (last release is now a bit
dated)

- Is it considered stable for production use?

- Does specifying multiple network interface in ggaoed.conf introduce
some kind of redundancy? (bd0 exported on eth0 and eth1 is still
available if eth0 goes down)

- Does specifying multiple network interface in ggaoed.conf introduce
some kind of throughput aggregation? (bdO export on eth0 and eth1 can
benefit from throughput of eth0 + eth1)