From mboxrd@z Thu Jan 1 00:00:00 1970 From: Steve Dickson Subject: Re: [PATCH] Reinstantiating stale inodes Date: Fri, 23 Apr 2004 12:16:11 -0400 Sender: nfs-admin@lists.sourceforge.net Message-ID: <4089414B.8090102@RedHat.com> References: <482A3FA0050D21419C269D13989C61130435DF4D@lavender-fe.eng.netapp.com> <1082732429.2617.35.camel@lade.trondhjem.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Cc: Charles Lever , nfs@lists.sourceforge.net Return-path: Received: from sc8-sf-mx2-b.sourceforge.net ([10.3.1.12] helo=sc8-sf-mx2.sourceforge.net) by sc8-sf-list2.sourceforge.net with esmtp (Exim 4.30) id 1BH3LM-0006kj-Bk for nfs@lists.sourceforge.net; Fri, 23 Apr 2004 09:16:00 -0700 Received: from mx1.redhat.com ([66.187.233.31]) by sc8-sf-mx2.sourceforge.net with esmtp (TLSv1:AES256-SHA:256) (Exim 4.30) id 1BH3LL-0005Ww-Uh for nfs@lists.sourceforge.net; Fri, 23 Apr 2004 09:16:00 -0700 To: Trond Myklebust In-Reply-To: <1082732429.2617.35.camel@lade.trondhjem.org> Errors-To: nfs-admin@lists.sourceforge.net List-Unsubscribe: , List-Id: Discussion of NFS under Linux development, interoperability, and testing. List-Post: List-Help: List-Subscribe: , List-Archive: Trond Myklebust wrote: >On Fri, 2004-04-23 at 10:48, Lever, Charles wrote: > > >>there appears to be a real problem when restoring from a backup, >>or using rsync. the file size and the mtime stay precisely the >>same, but the file handle changes. i'm not sure anything can be >>done about this in NFSv2/3? >> >> > >The only way to distinguish the two is to replace the use of the mtime >with the ctime in nfs_check_verifier() and friends. > > Again.. this was the first place I looked... And I just can't figure how to use NFS_CACHE_CTIME() in any meaningful calculation... >Under ordinary circumstances, the mtime and ctime should be more or less >identical, > In this particular case they aren't because the server explicitly maintains the "old" mtime and since mtime does not change we'll never get the new ctime... >so I'm not sure why Steve was seeing extra revalidations when >he was running the Connectathon suite. Were you perhaps changing the >algorithm in nfs_refresh_inode() instead, Steve? > > A very keen observation... Yes, when I introduced ctime into the valid or not valid calculation, the getattrs (i.e revalidations) but even more surprising (at least to me) was the number of reads sky rocked when I ran the cthon tests... SteveD. ------------------------------------------------------- This SF.net email is sponsored by: The Robotic Monkeys at ThinkGeek For a limited time only, get FREE Ground shipping on all orders of $35 or more. Hurry up and shop folks, this offer expires April 30th! http://www.thinkgeek.com/freeshipping/?cpg=12297 _______________________________________________ NFS maillist - NFS@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nfs