At 02:56 PM 2001-11-18, Thomas M. Root wrote:
>I'm in the process of migrating our 3000s from local disk drives to an EMC
>symmetrix. I won't change the production system until after the first of
>the year, but I plan to change the test system this week. I haven't done a
>migration like this since I used a Series II in 1979. The concept looks the
>same, but I'd like some advice on the 'new' details.
>
>The test system currently has a mix of 1.3G, 2G, and 4G drives (20 total)
>all in a single system volume set. I,m going to replace all the current
>drives with 10 logical 4G drives; 2 in the system volume set and 8 in a user
>volume set. I've run BULDACCT and done a full system backup and verify.
>
>My plan is to make a CSLT with a store of everything I want to place on the
>new system volumes. Then, I'll shut the system down, remove the current
>drives and controllers, connect the new controller and drives, and do an
>INSTALL from my CSLT. When the system comes up, I'll stream the BULDJOBs,
>restore the files from the CSLT to my system volume, build the remaining
>accounts on my user volume, and restore the user files from my backup.
>
>1. Am I on the right track here?
Yes - except you should not need to stream any BULDJOBs - in fact it's much
better to NOT do so. Instead, after doing the INSTALL, you simply restore
the directories that you had earlier stored onto your full backup.
The order of events should be as follows:
1. INSTALL from CSLT onto new system disk ldev 1.
2. use sysgen to configure non-ldev1 disk drives.
3. start norecovery
4. volutil to expand MPEXL_SYSTEM_VOLUME_SET and build and expand
uservolumeset.
5. restore directories from earlier full system backup (that contained the
;DIRECTORY;ONVS=MPEXL_SYSTEM_VOLUME_SET,uservolumeset options)
6. restore *t;/;create;show=offline;olddate;partdb;keep from the full backup
7. start norecovery again
8. make another cslt
>2. Can I configure the new drives during the INSTALL or should I do that
>when I make the CSLT?
See step 2 above
>3. Am I correct that if things don't work, I can put my original controllers
>and drives back, boot from the appropriate path, and have everything back as
>it was before? (ie 'Live to try another another day')
Yup. You have not burned any bridges.
But, you really shouldn't need to go back.
>4. Can I initialize and label 'new' disks as part of the INSTALL or do I
>need to use a running system to create scratch disks?
You can do it right after the INSTALL, using VOLUTIL. See step 2 above
>5. Have I left out anything?
Not that I can think of.
>In a seperate, but related issue, my production system has 7 2G drives in a
>HAAS box that is connected to a single ended SCSI controller. DSTAT All
>shows all the drives having SCRATCH status. Can I power down the box and
>unplug the cable from the controller without shutting down the system?
Yes, I don't see why not . Since you say that you have 7 devices on the
single-ended SCSI card, there cannot be any other devices attached to that
card.
>Thanks,
You're welcome.
>Thomas M. Root
>Desert Schools Federal Credit Union
>[log in to unmask] <home>
>[log in to unmask] <work>
>
>* To join/leave the list, search archives, change list settings, *
>* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
---------------------------------------------------------------------------
Gilles Schipper
GSA Inc.
HP System Administration Specialists
300 John Street, Box 87651 Thornhill, ON Canada L3T 7R4
Voice: 905.889.3000 Fax: 905.889.3001
Internet: [log in to unmask]
---------------------------------------------------------------------------
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
|
