From mboxrd@z Thu Jan 1 00:00:00 1970 From: Anton Eliasson Subject: Re: Broken nilfs2 filesystem Date: Wed, 12 Jun 2013 22:12:25 +0200 Message-ID: <51B8D629.4070301@antoneliasson.se> References: <1369809573.2253.31.camel@slavad-ubuntu> <20130529.233757.27789741.konishi.ryusuke@lab.ntt.co.jp> <1369894385.2255.8.camel@slavad-ubuntu> <20130530.155543.480320022.konishi.ryusuke@lab.ntt.co.jp> <1370501769.2202.36.camel@slavad-ubuntu> <20130606092054.GA201@diablo.13thmonkey.org> Mime-Version: 1.0 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <20130606092054.GA201-HNv6YvNvQKMNqjISwOrxaLFspR4gePGN@public.gmane.org> Sender: linux-nilfs-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-ID: Content-Type: text/plain; charset="iso-8859-1"; format="flowed" To: Reinoud Zandijk Cc: linux-nilfs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org Reinoud Zandijk skrev 2013-06-06 11:20: > Hi, > > just my $0.02 so to say: > > On Thu, Jun 06, 2013 at 10:56:09AM +0400, Vyacheslav Dubeyko wrote: >> First of all, unfortunately, I can't reproduce the issue yet, curren= tly. >> I suspect that in this issue the aging state of volume, peculiarity = of >> workload and environment play very important role. As I remember, al= l >> reporters of likewise symptoms (broken bnode error messages) talked >> about several months of successful working of NILFS2 file system. > sounds to me as if a b-tree is in a perculiar state and that updating= the > btree results in this corruption. > > Have you tried to mount one of the checkpoints/snapshots earlier as R= O and see > if those are correct? If so, dumping both DATs and both btrees might = give a > clue as to what went wrong. If only it gives a clue as to how complic= ated the > btree is before the updating and what actions are taken on it. > > With regards, > Reinoud > I have configured nilfs_cleanerd.conf to clean very aggressively so my=20 earliest checkpoint is from after the incident. I included the contents= =20 of that file in my first email sent on May 22=20 (http://article.gmane.org/gmane.comp.file-systems.nilfs.user/2920). Even so, I tried to loopback mount the oldest checkpoint I have which I= =20 found was affected by the same corruption. # losetup /dev/loop0 /Athena/Dump/riven/riven-home-20130531.img # mount /dev/loop0 /mnt $ mount | tail -1 /dev/loop0 on /mnt type nilfs2 (ro,relatime,norecovery) $ lscp /dev/loop0 CNO DATE TIME MODE FLG NBLKINC ICN= T 1260571 2013-05-23 16:51:49 cp - 140 15549= 6 1260572 2013-05-23 16:51:51 cp - 1632 15549= 5 1260575 2013-05-23 16:52:06 cp - 1473 15549= 6 1260576 2013-05-23 16:52:09 cp - 49 15549= 5 1260580 2013-05-24 23:36:11 cp - 1345 15549= 6 1260581 2013-05-24 23:36:16 cp - 1500 15549= 5 1260582 2013-05-24 23:36:21 cp - 1356 15549= 7 1260583 2013-05-24 23:36:26 cp - 1465 15549= 5 # chcp ss /dev/loop0 1260571 # umount /mnt # mount -o ro,norecovery,cp=3D1260571 /dev/loop0 /mnt $ cd /mnt/anton/Bilder/20130321-28\ Jakobs\ bilder\ fr=E5n\ Nederl=E4nd= erna $ LANG=3DC cat *>/dev/null cat: 160.JPG: Input/output error cat: 163.JPG: Input/output error cat: 164.JPG: Input/output error cat: 165.JPG: Input/output error cat: 170.JPG: Input/output error cat: 172.JPG: Input/output error cat: 179.JPG: Input/output error --=20 Best Regards, Anton Eliasson -- To unsubscribe from this list: send the line "unsubscribe linux-nilfs" = in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html