[Pvfs2-developers] bmi multiple address endpoints

[Sam Lang]

On Nov 9, 2007, at 11:11 AM, Pete Wyckoff wrote:

> slang at mcs.anl.gov wrote on Fri, 09 Nov 2007 10:57 -0600:
>> I was actually thinking the tabfile list would be optional, and  
>> its use
>> would be only to filter out addresses in the list provided for  
>> each Alias.
>> The BMI methods would still be enabled based on what was specified  
>> in the
>> BMIModules config option.  This still hurts performance because of  
>> the
>> sequential polling with multiple methods enabled, but I was  
>> thinking that
>> could be solved separately (multiple threads, etc.).  Should we  
>> try to not
>> enable methods specified in BMIModules if they're filtered out  
>> with the
>> mntent option?
>
> Oh, I see.  That would work too.
>
> As a client goes to send to a new address, it will auto-enable that
> method.  (Right?  It needs to do that to talk to the config server
> before it even sees fs.conf.)

Yes.  It auto-enables any method for an address that isn't already  
enabled.

>   Ditto for a server listening on an
> address.  So the BMIModules listing isn't really required, as far as
> I can tell.

Its not required by the clients, no.  And it does seem redundant on  
the server, esp. if we decide to specify options to the server that  
are exactly the endpoints/protocols to listen on.
-sam

>
> Yeah, for simplicity, you can just filter out addresses.  We should
> get around to the performance fix someday too.  I just happen to
> think all these other issues that have arisen should not get in the
> way of the failover feature.
>
> 		-- Pete
>