faulted zpool , do not resilver

Jean-Yves Avenard jyavenard at gmail.com
Sun Jan 9 05:47:47 UTC 2011


On 9 January 2011 16:08, Adam Vande More <amvandemore at gmail.com> wrote:
> On Sat, Jan 8, 2011 at 10:49 PM, Jean-Yves Avenard <jyavenard at gmail.com>
> wrote:
>> Any pointers on what I should do now?
>> All my data seems fine :(
> The web page mentioned in the error message contains further info, have you
> reviewed that? (argh, I see it's behind a login now, that's a recent
> change).  This should be it's equivalent:
> http://dlc.sun.com/osol/docs/content/ZFSADMIN/gbbwl.html

I did.. I can't find what this metadata: bit referes to though

> One way the meta data corruption can occur is the hard drive lies to ZFS
> when it issues a cache flush.  If the device says it did and in reality it
> didn't and there was a power outage this type of corruption can occur.
> http://mail.opensolaris.org/pipermail/zfs-discuss/2010-January/035740.html

I've been trying for 3 days to resolve those issues now.. I'm going to
just destroy the zpool ; I have a full mirror of that zpool on another

It's going to be the easiest way.

I'm still worried that for the first time I tried replaced disks, it
failed so miserably...

For some reason moving my SATA disk from on sata interface to another
failed. zpool would want to read the pool anymore, complaining that
the zpool version was too high (it's version 14, definitely ok).
Moving the disks bak to the original sata interface, and I can read
the pool just fine.

So What I did is unplug the diisk from the old interface, zpool import
it (now in degraded mode) and plug back the disk and wait for the
resilvering to complete.


More information about the freebsd-stable mailing list