On Mon, 2007-05-07 at 09:02 -0500, Chris Johns wrote:
> Paul,
> 
> You're correct, and this is a bug in recent kernels that is also now 
> present in RHEL5 (and Fedora Core 5, for that matter). We noticed the 
> bug a while ago when we started using the beta version of RHEL5, since 
> our product makes use of a mix of read-only and read-write mounts of 
> different directories within a single filesystem from the same NFS server.
> 
> The bug is not even restricted to a single export entry. We have, in the 
> simples example, 2 exports: an export of /export/foo/ro (read-only from 
> the server) and /export/foo/rw (read-write from the server). When our 
> RHEL5 client mounts /export/foo/ro first, read-only on the client side, 
> it gets mounted read-only. Now, however, when we mount /export/foo/rw 
> read-write on the client side, the client NFS code downgrades the mount 
> to read-only, to match the first mount from the server's filesystem.
> 
> A suggested workaround was to use a unique 'fsid' value for each export 
> entry on the server, thus faking the client out by making it believe 
> that each mount was from a different filesystem. While this would work, 
> it's an administration nightmare, particularly with a large 
> installation, or when using a NetApp filer, for instance.
> 
> The Red Hat bug that's currently open against this is 
> https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=209964.
> 
> On 11/20/06, Trond Myklebust (NFS developer) replied to a posting on 
> [EMAIL PROTECTED] about this:
> 
>     All,
>     > 
>     > Any insights on this?
>     > 
>     > I did find the following post from Trond which appears
>     > 
>     >    http://lkml.org/lkml/2006/10/18/264
>     > 
>     > Which seems to imply that this is a 'feature' and not
>     > a 'bug'?  Is that true - or have I missed something?
>       
> 
> 
>     That is quite correct (thanks for taking the time to search the list
>     archives). For various reasons, we're trying to tighten the NFS cache
>     consistency. That means that practices that were previously allowed
>     (such as mounting different directories from the same filesystem using
>     different mount options) have been restricted.
>     The ability to have the same filesystem mounted both read-only and
>     read-write will hopefully soon be allowed again, but it needs to be
>     fixed at the VFS level. The patches to do so have circulated for quite
>     some time now, but have not been merged due to disagreements over how
>     the user interface for 'mount --bind' should be in this case.

>From what I've found about 50% or so of the patches were accepted in
2.6.19. We need to fire up the discussion on this again because I can't
see any way forward until the rest of these patch set gets into the
mainstream kernel. I'll try but I'm not sure how far I'll get.

Ian


_______________________________________________
rhelv5-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/rhelv5-list

Reply via email to