[stgt] Quick Failover & Scalability questions

[FUJITA Tomonori]

On Wed, 13 Oct 2010 10:26:25 -0600
"Mark Lehrer" <mark at knm.org> wrote:

> I have a few tgtd servers running and they are doing very well.  I would 
> like to add a 2nd server as a backup so I have one question.  Assuming that 
> the block device mirroring and IP/Mac failover are perfect ( kind of like 
> assuming a frictionless surface in physics :) ), how well does stgt fail 
> over to a 2nd server?

I think that there are people that use such configuration.


> The simplest would be for the clients to reconnect to the new server and 
> re-establish communications.  However, how painful would it be for the new 
> server to keep the same sockets open for a truly seamless failover?  Again, 

I don't think it matters. The simplest solution (multipath) works well
for lots of people. Sharing the state between two hosts is too
complicated.


> I am only concerned about the tgtd internal states at this point - assume 
> that the block device mirroring as well as the 
> keepalived/heartbeat/iptables/fencing/etc issues are handled already (though 
> there would obviously be a good bit of integration work there!).
> 
> My 2nd question - I am starting to play with things like ionice and 
> scalability.  One critical aspect of scalability is limiting the first 
> connection's performance so it doesn't degrade so quickly as more clients 
> come online.  It appears that ionice will probably not have the features I 
> need; is this on the tgtd roadmap?  Even something as basic as a generic "io 
> operations per second" would be a good start.

We've just aded the feature to get such stats.
--
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