>Subject: Re: Disksuite question
>Date: 10/9/03 4:51 AM Hawaiian Standard Time
>> Once the system is built on the other U60, I figure then I can move to
>> want it. Question what is required? Can I simply just copy the contents of
>> /etc/lvm and the md part of the /etc/system. I'm pretty sure that if I run
>> metainit it will redo everything, and that would take a fair bit of time.
>> I'm wondering if there's a way to fake that out, and get the raid partition
>> installed without running metainit..
>Your're not delusional, we've done similar things here before. You
>generally can't copy the files exactly because you might get different
>controller or target numbers. But, as long as you are carefull you can
>use the information from the md.cf file to make a new md.tab file for the
>new system. A few things to be wary of:
>-When using metainit make sure to use the -k flag to avoid hosing your
Ok. I was wondering how to get metainit just to setup the /dev entry without
intializing the metadevice.
Note to self: Next time read the entire man page.
Quote:>-Make sure that the interlace is exactly the same and that you are SURE
>you have got the order of the disks in the raid correct.
Ok. Easy enough.
Quote:>-Make sure the individual disks work in the new system and are visible to
>the OS before attempting to add them to SDS.
Sounds good.. pardon my stupidity.. but what is SDS? Solaris DiskSuite?
Quote:>-Don't have any metadb replica's on the raid5 slices.
I didn't know you could do that. I've always put my metadb replicas on a
partition separate from the data partition. Either that or I'm very confused
>Other than that, you should be fine. Good luck!
So does this make sense:
1. Set up Raid 5 partition on off line machine
2. Attach Raid5 to new machine.
Make sure all disks are visible. Note disk names.
3. Add replicas to database.
4. Set up md.tab file with correct partition names and interlace size.
5. metainit -k d0 (or whatever device name you use)
7. rock n' roll
Dr. Stuart A. Weinstein
Ewa Beach Institute of Tectonics
"To err is human, but to really foul things up
requires a creationist"