[stgt] need some clarification regarding STGT

[Richard Sharpe]

On Fri, Dec 30, 2011 at 2:54 AM, ronnie sahlberg
<ronniesahlberg at gmail.com> wrote:
> I wouldn't worry too much about that old political document. Politics is boring.

Well, it might be boring, but there is plenty of it in some open
source arenas :-)

> There are many iSCSI targets for linux. STGT is one, SCST is another,
> and the one everyone thinks is the best right now is Lio.
> So you have plenty top pick from.

It is correct that there are many. I started with STGT but migrated to SCST.

> Some benefits of STGT for some use cases is that it runs in userspace.
> Other implementations have other benefits.

Well, SCST also allows you to write device handlers in user-space,
which is most likely the key requirement.

> STGT also have semi-mature implementations of VTL and Jukebox. I have
> no idea if LIO or SCST supports these. Maybe they do, and maybe they
> are even better than STGT and then everyone wins ! Who knows?

I disagree with some aspects of the VTL implementation (esp the mixing
of metadata with data) but I am sure that it could be ported to SCST
without too much trouble. I think there is a Jukebox implementation
for SCST as well.

SCST has scst_local and LIO has lio_loop, which both allow you to
access device handlers locally for testing and other interesting
purposes.

> Pick the implementation that suits you best.

This is a good piece of advice.

>> 1. Kernel side FILEIO
>> 2. Kernel side BLOCKIO
>> 3. User space FILEIO
>> 4. O_DIRECT FILEIO
>> 5. Async FILEIO
>> 6. Failover Clustering
>> 7. Bidirectional Commands
>> 8. Different threading models to choose the best performing

Bing or Google is your friend. You really should do your homework yourself.

-- 
Regards,
Richard Sharpe
--
To unsubscribe from this list: send the line "unsubscribe stgt" in
the body of a message to majordomo at vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html