[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-nfsv4
Subject:    Re: [PATCH 0/4] Series short description
From:       Steve Dickson <SteveD () redhat ! com>
Date:       2009-09-29 14:29:26
Message-ID: 4AC219C6.3090208 () RedHat ! com
[Download RAW message or body]



On 09/26/2009 10:07 PM, Chuck Lever wrote:
> I thought I should put my money where my mouth is, and post an
> implementation I would find acceptable, if only for discussion.
> This series has had some light testing with Solaris and Linux.
> 
> Some notes:
> 
> 1.  Trying the NFSv4 mount first means we hit basically the same hang that
>     Ian Kent has been chasing with autofs, if the server isn't responding.
>     I think a fg mount is going to hang anyway in this case, so this may
>     not be a show-stopper.
> 
>     Eventually the mount will either fail or succeed, but it will be after
>     several minutes instead of 1 minute which is the default retry setting.
> 
> 2.  The Linux NFS server I have seems to export shares over NFSv4 with a
>     different path than it does for NFSv2/v3.  I assume that's what your
>     pseudoroot patches are meant to address.
> 
>     But what do we do about Linux servers running older mountd's?  Trying
>     NFSv4 first means an autonegotiated mount fails in that case because
>     the NFSv4 share path is not the same as the NFSv2/v3 path.  Do we try
>     rewriting the mount path if we get ENOENT from the first attempt?
> 
> [cel@matisse ~]$ showmount -e seurat
> Export list for seurat:
> /export *
> [cel@matisse ~]$ sudo mount -t nfs seurat:/export /mnt
> [cel@matisse ~]$ sudo umount /mnt
> [cel@matisse ~]$ sudo mount -t nfs4 seurat:/export /mnt
> mount.nfs4: mounting seurat:/export failed, reason given by server:
>   No such file or directory
> [cel@matisse ~]$ sudo mount -t nfs4 seurat:/ /mnt
> [cel@matisse ~]$ sudo umount /mnt
> [cel@matisse ~]$ 
> 
> 3.  The version negotiation logic in my patches is straightforward enough
>     that it may allow hooks so that the exact negotiation behavior can be
Ok.. In the end, these patches function exactly as the previous purposed patches
with the only exception of the kernel version check that makes mount.nfs 
backward completable, which is a nice touch.....

So to put an end this discussion, I will commit you patches... 

Congrats.... and nice work!

steved.
   
_______________________________________________
NFSv4 mailing list
NFSv4@linux-nfs.org
http://linux-nfs.org/cgi-bin/mailman/listinfo/nfsv4
[prev in list] [next in list] [prev in thread] [next in thread]