ZFS root on MB Intel S3420GP

[Lystopad Olexandr]

 Hello, Jeremy Chadwick!

On Sun, Apr 17, 2011 at 11:49:12AM -0700
freebsd at jdc.parodius.com wrote about "Re: ZFS root on MB Intel S3420GP":
> On Sun, Apr 17, 2011 at 08:14:40PM +0400, Lystopad Olexandr wrote:
> > On Sun, Apr 17, 2011 at 04:54:34PM +0300
> > gkontos.mail at gmail.com wrote about "Re: ZFS root on MB Intel S3420GP":
> > > There is a nice guide in the WIKI regarding how to install your system with
> > > ZFS on root.
> > > 
> > > http://wiki.freebsd.org/RootOnZFS/GPTZFSBoot/RAIDZ2
> > > 
> > > In any case don't configure the raid in your controller and let ZFS take
> > > care of this.
> > 
> > Thanks for answer!
> > 
> > Is it possible to hot change disks with zfs raid on my motherboard?
> 
> This has little to do with ZFS and more to do with SATA.  You will need
> a hot-swap backplane for this to be possible.  Decent server chassis
> usually provide this.  We use Supermicro systems with hot-swap backplanes
> and they work fantastic with FreeBSD + ahci.ko.
> 
> If you do not have a hot-swap backplane, there is a very good chance
> "strange things" will happen when you yank power or the signal cable.
> I've personally tried it on a test system without a hot-swap bay.  When
> I pulled the SATA power connector from the hard disk, I saw a blue spark
> near the power connector and the entire system lost power.
> 
> I've blogged about hot-swapping SATA disks on FreeBSD with ZFS in use
> and with ahci.ko, with full kernel output and all necessary details:
> 
> http://koitsu.wordpress.com/2010/07/22/freebsd-and-zfs-hot-swapping-sata-disks-with-ahci/
> 
> Please note the blog post demonstrated how I went about upgrading disks
> without needing to power the system off.  Readers have commented how I
> could have done it all by using the spare bay I had, but I explicitly
> chose *not* to use that bay for the benefit of the readers who might not
> have a spare bay.
> 
> Furthermore, the "zpool offline" steps probably aren't needed (ZFS
> should note the disk as UNAVAIL immediately and the array should become
> degraded), same with "zpool online".  I should really refine those
> procedures, or re-do the post for present-day 8.2-RELEASE.
> 
> When doing administrative/maintenance tasks, I tend to do as much
> possible to ensure the kernel/system knows what I'm about to do.  :-)
> 
> If you want me to perform an actual disk failure (literally yanking a
> disk out of a bay while the disk is in use + part of a ZFS pool), I can
> do that without any worry and provide the results here.  Just ask.

Needed one reboot and change boot sequnce. Remote boy helps me.

I successfully migrate remotely from gmirror (ada0) to gpt+zfs(raidz1):

[root@ ~]# df -h
Filesystem                   Size    Used   Avail Capacity  Mounted on
zroot                        905G    569M    904G     0%    /
devfs                        1.0K    1.0K      0B   100%    /dev
zroot/tmp                    904G     35K    904G     0%    /tmp
zroot/usr                    906G    2.0G    904G     0%    /usr
zroot/usr/local              904G    142M    904G     0%    /usr/local
zroot/usr/local/pgsql        904G     33K    904G     0%    /usr/local/pgsql
zroot/usr/ports              904G     31K    904G     0%    /usr/ports
zroot/usr/ports/distfiles    904G     28K    904G     0%    /usr/ports/distfiles
zroot/usr/src                905G    242M    904G     0%    /usr/src
zroot/var                    904G    111M    904G     0%    /var
# zpool status
  pool: zroot
 state: ONLINE
status: One or more devices has experienced an unrecoverable error.  An
        attempt was made to correct the error.  Applications are unaffected.
action: Determine if the device needs to be replaced, and clear the errors
        using 'zpool clear' or replace the device with 'zpool replace'.
   see: http://www.sun.com/msg/ZFS-8000-9P
 scrub: resilver completed after 0h0m with 0 errors on Sun Apr 17 20:56:30 2011
config:

        NAME           STATE     READ WRITE CKSUM
        zroot          ONLINE       0     0     0
          raidz1       ONLINE       0     0     0
            gpt/disk1  ONLINE       0     0     5  12K resilvered
            gpt/disk2  ONLINE       0     0     0
            gpt/disk3  ONLINE       0     0     0

errors: No known data errors
[root@ ~]#

-- 
 Lystopad Olexandr