[libvirt] Re: iSCSI Multi-IQN (Libvirt Support)

Dave Allan dallan at redhat.com
Thu Oct 29 20:27:32 UTC 2009


Shyam_Iyer at Dell.com wrote:
> 
>> -----Original Message-----
>> From: Dave Allan [mailto:dallan at redhat.com]
>> Sent: Tuesday, October 27, 2009 2:57 AM
>> To: Iyer, Shyam
>> Cc: libvir-list at redhat.com; Bellad, Sudhir; Domsch, Matt; KM, Paniraja
>> Subject: Re: [libvirt] Re: iSCSI Multi-IQN (Libvirt Support)
>>
>>
> 
> <Snip>
>  
>> Ok, you're proposing what I thought you were proposing.  My objection
>> to
>> what you want to do is that I think this sort of complexity is better
>> done in the client application than in libvirt.  In other words, I
>> think
>> that *some code*, *somewhere* is going to have to loop through all the
>> sessions and create and delete each one and enumerate the LUs on each
>> session.  I think we are only debating whether that loop should be in
>> the client application or in libvirt.  Why do you think we should put
>> that complexity into libvirt?
>>
> 
> Ok. I was looking at giving an api that will facilitate this for the
> client easily but I see your point as well in the context of the present
> design..
> 
> So, we will stick to the following paradigm-
> 
> 1) No initiator iqn in the xml would mean use of the default initiator
> iqn name
> 2) Initiator iqn provided would mean a unique session to be created
> using the provided iqn name.
> 3) Single iSCSI session per pool

Sounds good--thanks.

> A few points though -
> 
> The virt-manager GUI will need more intelligence built around easy
> manageability of pools.
> 
> 1) More information on the pools.
> 2) Ability to group multiple pools.
> 2) Ability to delete multiple pools for easy administration.
> 
> 
> * This would be useful in an environment where there may be tens of
> arrays, hundred pools and a few more hundred LUNs.
> * For the visual experience if we stick to the 1session per pool design
> we would find multiple pools for the same target iqn and also pointing
> to the same luns and filing up the GUI screen unnecessarily sometimes
> ...

You should see the same LUNs multiple times if some LUNs are visible on 
multiple sessions, regardless of how the pool is implemented as far as 
one or multiple sessions.  Am I understanding correctly that you want 
some sort of uniqueness filter on LUNs?

> Should this be discussed in the virt-manager lists also ?

I think this should be discussed primarily on the virt-manager lists. 
If there's work required in libvirt to implement what you want to see in 
virt-manager, then discuss in both places, but you should get a design 
hammered out there and then see what may make sense to implement in libvirt.

Dave




More information about the libvir-list mailing list