Hi all,

I have an Ultra 2, Solaris 9, Generic_118558-11, with a D1000
attached. The ULT2 died so I moved the array to another Ultra 2, same
flashprom, same kernel rev. The D1000 had 2ea, RAID-5 arrays, 3 drives
each. I reused the same controller in the same slot.

I had previously saved the /etc/lvm files and /kernel/drv/md.conf
files just in case.

I had previously had this same problem 2 years ago, and was successful
moving the array to a new host. I'm having trouble getting metastat to
see the state database replicas after manually entering them in to /
kernel/drv/md.conf

I Followed the Solaris Volume Manager guide to move an array from one
host to another.

Original file info from Ultra 2 before it died:

/etc/md.cf

<-
d1 -r c1t0d0s6 c1t1d0s6 c1t5d0s6 -k -i 32b
d0 -r c1t2d0s7 c1t3d0s7 c1t4d0s7 -k -i 32b
->

/etc/mddb.cf

<-
sd 23 16 id1,sd@w00e09e00001f47e4/h -2065
sd 31 16 id1,sd@w00e09e0001b48f1e/h -2121
sd 39 16 id1,sd@w00e09e0001b9336d/h -2082
sd 6 16 id1,sd@w0010b9fe56e90200/g -2051
sd 14 16 id1,sd@w0010b9fe38eb0200/g -2100
sd 46 16 id1,sd@w0010b9feae5c0300/g -2177
->

/kernel/drv/md.conf

<-
mddb_bootlist1="sd:23:16:id1,sd@w00e09e00001f47e4/h sd:
31:16:id1,sd@w00e09e0001b48f1e/h sd:39:16:id1,sd@w00e09e0001b9336d/h s
d:6:16:id1,sd@w0010b9fe56e90200/g sd:14:16:id1,sd@w0010b9fe38eb0200/g
sd:46:16:id1,sd@w0010b9feae5c0300/g
->


From new Ultra 2 host to get major/minor numbers

ls -Ll /dev/dsk/c1t0d0s6 = 32/126
ls -Ll /dev/dsk/c1t1d0s6 = 32/134
ls -Ll /dev/dsk/c1t5d0s6 = 32/166

ls -Ll /dev/dsk/c1t2d0s7 = 32/143
ls -Ll /dev/dsk/c1t3d0s7 = 32/151
ls -Ll /dev/dsk/c1t4d0s7 = 32/159

so with the new minor device numer from the ls -Ll command and plugged
them into the /kernel/drv/md.conf file like:

<-
mddb_bootlist1="sd:143:16:id1,sd@w00e09e00001f47e4/h sd:
151:16:id1,sd@w00e09e0001b48f1e/h sd:159:16:id1,sd@w00e09e0001b9336d/
h sd:126:16:id1,sd@w0010b9fe56e90200/g sd:
134:16:id1,sd@w0010b9fe38eb0200/g sd:166:16:id1,sd@w0010b9feae5c0300/
g"; md_devid_d
estroy=1;
->

Then rebooted.

metastat should see the state database replicas but it says:

<-
there are no existing databases
->

What have I done wrong?
Can I use smc to find these old state database replica's? I did it a
couple years ago, and recovered the array data perfectly.

I appreciate any pointers.

Thanks in advance,

Willie

"wallymcdoogle" <willief@base-2.com> wrote in message
news:0aa344ea-ec2b-49e6-8f52-c7f4e08cea1d@59g2000hsb.googlegroups.com...
> d1 -r c1t0d0s6 c1t1d0s6 c1t5d0s6 -k -i 32b
> d0 -r c1t2d0s7 c1t3d0s7 c1t4d0s7 -k -i 32b

The following is probably what you need:

% metainit d0 -r c1t2d0s7 c1t3d0s7 c1t4d0s7 -k -i 32b
% metainit d1 -r c1t0d0s6 c1t1d0s6 c1t5d0s6 -k -i 32b

This should make your RAID 5 devices seen by the new computer.
The -k tells the metainit that data is there and not to zero out the
metadevice.

What about metadatabases? Do you have a slice on each disk to hold that?

Trinean