Hello,
I have an A1000 (under Solaris 9). I used RM6 to create an 11 disk
RAID5 array - one LUN/logical drive. I then used Solaris Volume
Manager to create a single Concat/Stripe (d4) on the LUN. I then
created multiple soft partitions on top of the Concat/Stripe (d40, d41,
etc). I create my file systems on top of the soft partitions. I wanted
the flexibility to create more file systems and to grow them as need
be. Not sure this was the best way, but now my problem is that suddenly
(after working for a while) my Concat/Stripe says it is "State:
Unavailable" even though all the Soft Partitions on top of the
Concat/Strip say they are "State: OK". I cannot mount the soft
partition (no suprise since the underlying Concat/Stripe is not
available). They thing is, there is nothing wrong (that I can tell)
with the A1000 LUN on top of with the Concat/Strip was placed. Is there
some way to fix or diagnose this better? Can I recover the soft
partitions and or make the Concat/Strip "Available"? I tried playing
with 'metarecover' with no success.

Sincerely,

- Henrik

henrik <henrik@stat.tamu.edu> wrote:
> They thing is, there is nothing wrong (that I can tell)
> with the A1000 LUN on top of with the Concat/Strip was placed. Is there
> some way to fix or diagnose this better? Can I recover the soft
> partitions and or make the Concat/Strip "Available"? I tried playing
> with 'metarecover' with no success.

Any errors in the messages file? I might want to do a
'format,analyze,read' on the LUN just to verify that it's okay.

Then I'd capture the output of 'metastat -p'. Next I'd reboot and see
if the metadevice would come up properly when initialized.

If not, I'd be tempted to destroy the devices (metaclear) and recreate
them in the exact same order (helped by the metastat -p output).
Creating concat/stripes, one way mirrors, and soft partitions should not
affect underlying data. Creating raid 5 will zero storage unless you
also use the -k option.

I'd probably want to make sure I had a good backup before doing that.

--
Darren Dunham ddunham@taos.com
Senior Technical Consultant TAOS http://www.taos.com/
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >

Thanks for the response. The following command fixed the problem:

metadevadm -u c2t5d0

on restart the Concat/Stripe (created on top of the hardware RAID5 on
the A1000) is OK again. All the underlying soft partitions can be
mounted. What exactly happened to cause the problem in the first place
(i.e. what caused the device ID to change) is a mystery to me and has
me somewhat concerend.

Sincerely,

- Henrik


Darren Dunham wrote:
> henrik <henrik@stat.tamu.edu> wrote:
> > They thing is, there is nothing wrong (that I can tell)
> > with the A1000 LUN on top of with the Concat/Strip was placed. Is there
> > some way to fix or diagnose this better? Can I recover the soft
> > partitions and or make the Concat/Strip "Available"? I tried playing
> > with 'metarecover' with no success.
>
> Any errors in the messages file? I might want to do a
> 'format,analyze,read' on the LUN just to verify that it's okay.
>
> Then I'd capture the output of 'metastat -p'. Next I'd reboot and see
> if the metadevice would come up properly when initialized.
>
> If not, I'd be tempted to destroy the devices (metaclear) and recreate
> them in the exact same order (helped by the metastat -p output).
> Creating concat/stripes, one way mirrors, and soft partitions should not
> affect underlying data. Creating raid 5 will zero storage unless you
> also use the -k option.
>
> I'd probably want to make sure I had a good backup before doing that.
>
> --
> Darren Dunham ddunham@taos.com
> Senior Technical Consultant TAOS http://www.taos.com/
> Got some Dr Pepper? San Francisco, CA bay area
> < This line left intentionally blank to confuse you. >

"henrik" <henrik@stat.tamu.edu> writes:

> Thanks for the response. The following command fixed the problem:
>
> metadevadm -u c2t5d0
>
> on restart the Concat/Stripe (created on top of the hardware RAID5 on
> the A1000) is OK again. All the underlying soft partitions can be
> mounted. What exactly happened to cause the problem in the first place
> (i.e. what caused the device ID to change) is a mystery to me and has
> me somewhat concerend.

I think we have (had?) a similar problem here (A1000 + Solaris 9) which we
solved (by-passed?) by postponing the mounting of the vfstab entries that
are on A1000 based soft partitions (don't forget to change "mount at boot"
option in /etc/vfstab to no!). There is probably some race between A1000
initialisation and the /etc/vfstab mounting processes.

Bye, Dragan

--
Dragan Cvetkovic,

To be or not to be is true. G. Boole No it isn't. L. E. J. Brouwer

!!! Sender/From address is bogus. Use reply-to one !!!