Hi, On Wed, Jul 03, 2002 at 02:21:07PM -0600, Andreas Dilger <adilger@turbolinux.com> wrote: > On Jul 03, 2002 20:05 +0100, Anton Altaparmakov wrote: > > I just noticed that my file server running 2.4.18-pre3 + IDE patches & > > NTFS patches has this error message in the logs: > > > > EXT3-fs error (device md(9,4)): ext3_free_blocks: Freeing blocks not in > > datazone - block = 33554432, count = 1 > > > > This is the only ext3 error I have seen and the uptime is currently over > > 74 days. The error actually appeared two weeks ago. The timing coincides > > well with when this device (/dev/md4, a MD RAID-1 array) ran out of > > space, so it may well be related. > > This was fixed in newer kernels. No, the out-of-space behaviour was either an ext3_error if we ran out of inodes, or a minor i_nrblocks accounting error if we ran out of disk blocks. The bad pointer here is 0x2000000, and a bit flip like that is more often an indicator of bad memory than anything else. Cheers, Stephen
