DragonFly BSD
DragonFly users List (threaded) for 2012-10
[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]

Re: Root mount failed:5

From: Matthew Dillon <dillon@xxxxxxxxxxxxxxxxxxxx>
Date: Fri, 12 Oct 2012 10:35:43 -0700 (PDT)

:The txt file is 20GB and when bzipped 1.3 GB.
:|MD5 (bd-show.txt.bz2) = 9727a126b93316dba194ff596e9661ff
:I am uploading it to leaf:/home/sgeorge/crash/2012-10-12/bd-show.txt.bz2
:Will take 5 hours from now to complete

    Ok, I see it.  Well... interesting.  It does not look like a live-dedup
    related issue, which makes sense since it shouldn't have been possible
    for you to turn on live-dedup.  I wonder if there's an end-run around
    that though if you overrode it in loader.conf.  But there's no way it
    would have been turned on by something in sysctl.conf.

    Basically the B-tree dump shows three errors.  Two B-Tree nodes have a
    self-CRC error.  One of the nodes is very close to the root which is
    why your mount is failing... It can't get to the root directory due to
    the bad B-Tree node.

    I'm beginning to suspect that it might be the 'ad' (IDE driver)
    attachment.  Even though its a SATA disk, there might be issues
    with the IDE compatibility mode implemented by the BIOS / hardware.

    If your BIOS has an 'AHCI' mode you want to enable it, for sure.
    What kind of machine is this stuff running on?

    It could also be some sort of rare software bug but it's hard to
    contemplate that because all of our production machines have been
    running HAMMER for ages and have no b-tree errors beyond some historical
    mirror_tid propagation warnings (which don't effect operation).

					Matthew Dillon 

[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]