On Thu, 2014-03-27 at 13:39 +0100, Piotr Szymaniak wrote: > [ 1109.889697] mmcblk0: mmc0:b368 SMI 7.51 GiB > [ 1109.891261] mmcblk0: p1 p2 p3 > [ 1129.467450] NILFS warning: mounting unchecked fs > [ 1129.956859] NILFS: recovery complete. > [ 1129.960064] segctord starting. Construction interval = 300 seconds, CP frequency < 30 seconds > [ 1129.979344] NILFS warning: mounting fs with errors > [ 1283.005874] attempt to access beyond end of device > [ 1283.005882] mmcblk0p3: rw=8192, want=145241775180196072, limit=15120384 > > ~ # nilfs-tune -l /dev/sdf3 > > > nilfs-tune 2.1.6 > > > Filesystem volume name: (none) > > > Filesystem UUID: 53760664-f9ed-4c8d-af42-c6ee2f16d956 > > > Filesystem magic number: 0x3434 > > > Filesystem revision #: 2.0 > > > Filesystem features: (none) > > > Filesystem state: invalid or mounted,error > > > Filesystem OS type: Linux > > > Block size: 4096 I suppose that you have corrupted file system. Because it is possible to see that file system has 4 KB block size but the read request is 8 KB in size (moreover, this request is out of device). You are using this NILFS2 volume from 2012 year. So, the reason of the issue can be only corrupted block on the volume. I assume that there is significant probability of occurrence the issue with competition of segments for dirty blocks during segment construction. I've reproduced this issue very stably when it takes place simultaneous checking for updates and some intensive activity with volume (for example, compilation). So, such issue results with corruption of file system. But you can discover such corruption far away from place of origin. Anyway, I'll happy to investigate your issue if you can describe a reproducing path. I need in stable reproducing of the issue for investigation and fix. Could you describe it? Thanks, Vyacheslav Dubeyko. -- To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
