From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: with ECARTIS (v1.0.0; list xfs); Mon, 16 Oct 2006 06:54:11 -0700 (PDT) Received: from cuda.sgi.com (cuda2.sgi.com [192.48.168.29]) by oss.sgi.com (8.12.10/8.12.10/SuSE Linux 0.7) with ESMTP id k9GDrxaG017924 for ; Mon, 16 Oct 2006 06:54:00 -0700 Received: from sandeen.net (sandeen.net [209.173.210.139]) by cuda.sgi.com (Spam Firewall) with ESMTP id 5ED3FD196C52 for ; Mon, 16 Oct 2006 06:53:15 -0700 (PDT) Message-ID: <45338EC9.9010501@sandeen.net> Date: Mon, 16 Oct 2006 08:53:13 -0500 From: Eric Sandeen MIME-Version: 1.0 Subject: Re: recover xfs with v1 dirs on linux? References: <2117b660610160232p1ffb28e0necfeeb377b5d4ead@mail.gmail.com> In-Reply-To: <2117b660610160232p1ffb28e0necfeeb377b5d4ead@mail.gmail.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: xfs-bounce@oss.sgi.com Errors-to: xfs-bounce@oss.sgi.com List-Id: xfs To: Honza Fikar Cc: xfs@oss.sgi.com Honza Fikar wrote: > Hi everybody, > > I want to recover old SGI Irix xfs disks, but I don't have any more the old > hardware. I was able to copy the xfs partition in a file with dd and I could > run xfs_recover, xfs_check and xfs_ncheck, which gives me the filenames. But > I was not able to mount this file. Kernel 2.6.17 gives "Function not > implemented", while 2.6.5 mounts, but the disk looks empty. > > I think the reason is "v1 dirs": > > file img4.xfs > img4.xfs: SGI XFS filesystem data (blksz 4096, inosz 256, v1 dirs) > > Is there a way to get the files? I have the innode numbers and filenames as > an output of xfs_ncheck: > .... > 947364 abc/MIXDUM/MD110/20K/step9/dynrt.out.gz > 184821 aa/Monteal/fmd3kev-45K/cluster.I.076.gz > 1286273 abc/MIXDUM/MD111/250K/step10/dyn5.f1.gz > ... v1dir support was removed a short while ago, they never really worked on Linux. http://oss.sgi.com/projects/xfs/faq.html#useirixxfs Older kernels had rudimentary v1 support, and you may find some degree of functionality, or, as it seems in your case, no functionality at all. If this is terribly important data, then some finite amount of work (involving a patched glibc, IIRC) might be able to get you going. -Eric