ext2 FS corruption with 2.5.59.

From: Oleg Drokin (green@namesys.com)
Date: Thu Jan 23 2003 - 07:38:32 EST


    While doing my usual quick tests I pass all my reiserfs patches through,
    I accidentally forgot to mount reiserfs partition and all the tests
    were performed on ext2 instead.
    So I found that first of all fsx (freebsd nfs testing tool)
    breaks if there are parallel writers (no matter if they write to this
    same partition or not). i_blocks value becomes negative on truncate
    (and then it is written to disk). fsx log attached.
    Also subsequent fsck found lots of other corruptions.
    (previously fs was checked and no corruptions were found).
    Excerpts from e2fsck output are attached.

    My test consists of running "fsx -c 1234 testfile", "iozone -a",
    "dbench 60", "fsstress -p10 -n1000000 -d ." at the same time on the
    tested FS.
    fsx usually breaks just when dbench is finished.

    Host is dual athlon 1700+, 1G RAM, SMP kernel, Highmem support, no preemt.

    I reproduced this behaviour on vanilla 2.5.59. (and It is easily

    "..." symbols in fsck output excerpts mean "lots of these".
angband:~ # /sbin/e2fsck -fn /dev/hda1 | wc -l
e2fsck 1.24a (02-Sep-2001)


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

This archive was generated by hypermail 2b29 : Thu Jan 23 2003 - 22:00:31 EST