From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S964932AbXC1RXN (ORCPT ); Wed, 28 Mar 2007 13:23:13 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S964862AbXC1RXM (ORCPT ); Wed, 28 Mar 2007 13:23:12 -0400 Received: from mail.fieldses.org ([66.93.2.214]:54244 "EHLO fieldses.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S964932AbXC1RXM (ORCPT ); Wed, 28 Mar 2007 13:23:12 -0400 Date: Wed, 28 Mar 2007 13:23:10 -0400 To: Phy Prabab Cc: linux-kernel@vger.kernel.org Subject: Re: Issues with knfsd + solaris clients Message-ID: <20070328172310.GC21221@fieldses.org> References: <2fe399760703271334j1fe56c9ctc83e1310620d7728@mail.gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <2fe399760703271334j1fe56c9ctc83e1310620d7728@mail.gmail.com> User-Agent: Mutt/1.5.13 (2006-08-11) From: "J. Bruce Fields" Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Mar 27, 2007 at 01:34:21PM -0700, Phy Prabab wrote: > Here is a little bit more information on my issue with slowlaris and Could you just use their name? It's a little distracting otherwise. > knfs 2.6.21-rc4/5 (actually appears in 2.4.20.x). Running this from > the client (x.org source code): > >ls -l xc/lib/Xmu/Xct.h > -rw-rw-r-- 1 phy bdf 0 2006-01-09 06:59 xc/lib/Xmu/Xct.h > >file xc/lib/Xmu/Xct.h > xc/lib/Xmu/Xct.h: cannot read: Is a directory > >rm -rf xc/lib/Xmu/Xct.h > rm: cannot remove `xc/lib/Xmu/Xct.h': Not owner > >uname -a > SunOS mm3 5.10 Generic_118855-14 i86pc i386 i86pc Solaris > >chmod 777 xc/lib/Xmu/Xct.h > >rm -rf xc/lib/Xmu/Xct.h > rm: cannot remove `xc/lib/Xmu/Xct.h': Not owner > > > > I would think the file would not be directory and if I tried to remove > it, I should be able as owner and permissions are correct. This is on > a solaris 10 on intel, but I also see this problem on solaris 8, 9 and > 10 on sparc. Again, AIX5303 and 04 and all linux clients (2.4.x, > 2.6.x) seem to be fine. Would it be possible to get a trace showing what the rpc traffic during the above test? You could run something like tcpdump -s0 -wTMP host my-client.example.com and then post a copy of TMP someplace. Or look at it yourself in wireshark--what you're looking for is how the server replied to the rpc's generated by the commands above, and whether the server's replies were consistent. --b.