Steve Peng wrote:

> Lars Tunkrans wrote:
>
>> The 146 GB disk are standard drives for the V240. I had done the following :
>>
>>  # metadb -a -f -c2  c1t0d0s7 c1t1d0s7 c1t2d0s7 c1t3d0s7
>>
>> and then setup  stripes and mirrors  with  metainit .  run metaroot  and 
>> fixed  up /etc/vfstab.   Then after  the reboot , the system could not mount 
>> the Root
>
>
> Ah!  I like to know how did you set up your root mirroring - step by step 
> from creating stripe.


The Four SCSI disk in the V240 is intended to be used as follows.

Boot disk  is  72 GB c1t0d0   
root mirror is 72 GB c1t1d0
Data disk 1  is 146 GB  c1t2d0
Data disk 2 is   146 GB c1t3d0

I followed what I belive is a rather normal procedure.

metainit  -f  d11  1  1  c1t0d0s0          root    10 GB
metainit  -f  d21  1  1  c1t0d0s1          var       5 GB
metainit  -f  d31  1  1  c1t0d0s3           swap    9 GB
metainit  -f  d41  1  1  c1t0d0s4           home   5 GB
metainit  -f  d51  1  1  c1t0d0s5          Application slice    40 GB
metainit  -f  d61  1  1  c1t2d0s0         This  slice was   the whole  146 GB  
disk apart from the 128 MB mddb at then end

 [ rootmirror disk ]
metainit  -f  d12  1  1  c1t1d0s0          root    10 GB
metainit  -f  d22  1  1  c1t1d0s1          var       5 GB
metainit  -f  d32  1  1  c1t1d0s3           swap    9 GB
metainit  -f  d42  1  1  c1t1d0s4           home   5 GB
metainit  -f  d52  1  1  c1t1d0s5          Application slice    40 GB
metainit  -f  d62  1  1  c1t3d0s0         This  slice was   the whole  146 GB  
disk apart from the 128 MB mddb at then end


metainit  d10 -m d11
metainit  d20 -m d21
metainit  d30 -m d31
metainit  d40 -m d41
metainit  d50 -m d51
metainit  d60 -m d61

metaroot  d10

Then checked  /etc/system   and edited up  /etc/vfstab  with md  devicenames.

# sync;sync
# init 6

Anyways, I dont think  mysetup mattered - in the reboot  all mddb's  were lost  
and  the "d10"
root device could not be found. I will do some more investigations next week if 
shedule allows.
What I'll do first  Is to  put the mddb's  first on the 146 GB disks  and see 
if they survives a reboot.
As the mddb's  are permanent if I only place them on the 72 GB disks I' ll go 
with that solution if I
have to.
This  should really be tried on additional systems , any volunteers ?
//Lars
This message posted from opensolaris.org

Reply via email to