From: Chuck Lever Subject: Re: NFS: unknown mount option: grpid Date: Thu, 10 Apr 2008 11:17:48 -0400 Message-ID: References: <47FE0C42.6030900@TechFak.Uni-Bielefeld.DE> <56D1CF7A-BB26-4D07-B47D-9D43D55A1B8F@oracle.com> <47FE2C2D.1070703@redhat.com> Mime-Version: 1.0 (Apple Message framework v753) Content-Type: text/plain; charset="us-ascii" Cc: NFS list , Jan Sanders , Linux NFSv4 mailing list To: Peter Staubach Return-path: In-Reply-To: <47FE2C2D.1070703@redhat.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: nfsv4-bounces@linux-nfs.org Errors-To: nfsv4-bounces@linux-nfs.org List-ID: On Apr 10, 2008, at 11:03 AM, Peter Staubach wrote: > Chuck Lever wrote: >> Hi Jan- >> >> On Apr 10, 2008, at 8:46 AM, Jan Sanders wrote: >> >>> Hi all, >>> >>> I have come across a little NFS problem. >>> >>> My nfs client, a Ubuntu-Hardy machine with nfs-common-1.1.2 tries to >>> mount a directory but fails complaining >>> >>> Apr 10 12:18:34 sorpe kernel: [ 490.911951]NFS: unknown mount >>> option: grpid >>> >>> The mount options are rw,nosuid,grpid. The mount is done by >>> autofs but >>> trying to mount the directory usdin the same options rw,nosuid,grpid >>> results in the same error. >>> I checked using strace that the mount call was indeed done using >>> grpid. >>> The mount call returns with EINVAL invalis argument. >>> >>> From strace: >>> mount("nfs-server:/volumes/www", "/vol/www", "nfs", MS_NOSUID, >>> "grpid,addr=192.168.0.123") = -1 EINVAL (Invalid argument) >>> >> >> grpid isn't a valid NFS mount option; it's valid only for xfs and >> ext2/3, according to mount(8). >> >> I can't explain why the earlier version of mount.nfs didn't >> complain about it. >> >> > > I thought that the NFS mount command was supposed to ignore > mount options that it didn't understand. It could perhaps > give a warning message, but should mount anyway. > > I thought that this behavior was useful for automounter > applications which have to be able to share maps in a > heterogeneous environment. Well, it does ignore legacy NFS mount options that are no longer supported. However, I was not aware of a requirement for NFS mount to ignore all options it doesn't understand. It's easy enough to add, I suppose. Community opinion? -- Chuck Lever chuck[dot]lever[at]oracle[dot]com