Recovering a GEOM RAID0 array

[Juan Miscaro]

2009/3/31 hv <hv at tuebingen.mpg.de>:
>
> Am 30.03.2009 um 21:21 schrieb Juan Miscaro:
>
> Hi gang,
>
> I'm running a remote 6.2 system which recently got shut down
> unexpectedly (tower was physically nudged and apparently lost power).
> I am running a 2-disk striped array with the geom_stripe.ko module.
>
> So my fstab line is
>
> /dev/stripe/st0a        /data           ufs     rw,acls         2       2
>
> Thing is, the /dev/stripe directory no longer exists.
>
> Make sure the the geom_stripe module is actually loaded. (with kldstat)
> If it's not already loaded, do a kldload geom_stripe (which creates the
> /dev/stripe directory, if the metadata exist on the disk ("automatic mode")

It *was* in automatic mode since it's been running and rebooting for
the past 2 years without any commands being issued in scripts.

The module was loaded but I unloaded it and reloaded it.  Strangely,
the size of the module (as given by 'kldstat') changed.  The output
line went from

 2    1 0xc0b0c000 6c9c     geom_stripe.ko

to

 4    1 0xc4f42000 6000     geom_stripe.ko

The /dev/stripe directory is still not present.

> and perform a normal fsck afterwards. (e.g fsck -y /dev/stripe/st0a)
> If the metadata /(manual mode) are not stored on the disk, you need to issue
> exactly the same commands to recreate it as you originally did.
> All this is nicely documented in the man page (man gstripe)
> By adding the following line
> geom_stripe_load="YES"
> to /boot/loader.conf you make sure the stripe will be available after reboot
> (at least in "automatic mode").

Yes, the line is there (in /boot/loader.conf).

I fear the boot sector of one of the drives has been corrupted.

--
jm