[RFC,PATCH 0/4] Dynamic Pseudo Root

[Chuck Lever]

On Dec 10, 2007, at 4:46 PM, Neil Brown wrote:
> On Monday December 10, chuck.lever at oracle.com wrote:
>>> Question: How often are do people run multiple rpc.mountd daemons  
>>> and
>>> for what reason?
>>
>>
>> I don't know the answer to that question, but the mount command has
>> options to use different RPC versions and program numbers when
>> contacting the mountd on an NFS server, ostensibly to allow server
>> admins to provide unique NFS services on the same server.
>
> I had alway assumed that was to allow multiple implementations of nfs.
> e.g. you could be using knfsd and the rest of nfs-utils to export some
> filesystems, and unfsd (or some other user-space nfsd) to export some
> other data.
>
> I cannot see any way to usefully run two independent instances of the
> nfs-utils mountd on the same host....

While implementing text-based mount option parsing, I wondered if it  
was worth it to continue supporting some of these legacy options.   
Some features of mount (like "bg" mounts) have been broken for quite  
a while, and no-one has complained.

I didn't implement support for nfsprog in the kernel parsing  
routines, for example, since the in-kernel NFS client hasn't  
supported using a different RPC program number for NFS for a very  
long time.

> Hmmm.  I wonder if the current multi-virtual-namespace craze is going
> to hit nfsd and want us to support multiple independent nfsds on the
> one host (presumably with separate IP addresses etc).  I guess we
> could arrange that if the nfsd filesystem were mounted in different
> places (or in different namespaces) it provided independent caches
> .... I think I'll decide that is someone else's problem :-)


Thanks for the chuckle :-)

--
Chuck Lever
chuck[dot]lever[at]oracle[dot]com