DL380G9 SA 440ar - Howto clone/move [boot] RAID1 between boxes

[SysAdmin2]

Want to clone installation from one server to another, finding many others having issues with moving raid sets between servers (eg. Moving-RAID-to-another-Proliant) which talks of a document on subject:

RAID 1(+0): breaking mirrors and rebuilding drives http://h20000.www2.hp.com/bc/docs/support/SupportManual/c00378986/c00378986.pdf

only this link seems dead. Same with solution link in this question, also dead.

Any one knows why to find newest doc on subject or howto eg. clone a boot raid1 set from one server to another by moving only one side of mirror to a new identical server (pick up raid1 set and resync to new fresh 2.side drive)?

When I try to move target server wouldn't enable logical array found on drive. Would also like to be able to simply move all drives to a new peice of HW in case of eg. a board failure and then pick up logical arrays and continue on new HW. 

[Jimmy Vance]

Link to the document is http://h20565.www2.hpe.com/hpsc/doc/public/display?docId=emr_na-c00378986  Many of the documents you can change them from hp.com to hpe.com and they will work again.

As to your question. The RAID information is stored on each drive. You can move all the drives from one system to another without issue.  To clone a RAID 1, I use the manual method shown in the guide.  Make sure you shut down the server before removing the drive, and the system you are moving to needs to be at the same or newer array firmware

[SysAdmin2]

@Jimmy thanks but your link also returns:

'We are sorry but your search produced 0 results'

What's the name of the doc you ment to refer to (then I might be able to search it out of the vast HPE content)?

[Jimmy Vance]

How odd, as I see the error now, but wasn't getting it when I copied the link earlier

Simple process, fFrom what I was able to pull up earlier

Breaking a mirror and rebuilding the array using the manual method

---

CAUTION Use the following procedure for configurations with only one array on the controller. If you configure multiple arrays on a single controller, all arrays must be broken and restored together. Because of the added complexity for multiple arrays, we discourage this method of creating a backup of the logical driver.

---

You can break a RAID 1 or RAID 1(+0) mirror simply by removing a drive while the server is online. You must shut down the server first if you want to preserve the data on drives you are removing. If you want to use the removed drives to restore the server, or if you need the data on the drives for any other reason, follow these steps exactly to break a mirror:

1. Verify that the controller contains RAID 1 or RAID 1(+0) logical drives only.
2. Shut down the OS to ensure a clean file system shutdown.
3. Power off the server.
4. Remove one-half of the drives created by the mirror (the backup). Label each removed drive with its respective drive bay number.

---

NOTE For two-drive configurations, just remove one drive. For larger RAID 1+0 configurations, the first n/2 drives are always mirrored to the second n/2 drives (unless drives have been physically reordered using Offline Drive Movement). Always remove the set of drives that are one side of the mirrored array. Do not remove drives from opposite sides of the mirror.

---

5. Restart the OS with the remaining half of the drives that the mirror created.
6. You will get a message during boot-up indicating that the removed drives are not responding (Figure 1).
7. Press F2 to fail the missing drive(s) and continue in recovery mode. All drives on the controller will remain disabled until you press F2. If you press F1 by mistake, simply reboot again and press F2 to fail the drives.

If you are breaking the mirror to test an OS patch, you should normally test the OS patch before making the array fault-tolerant again. Using this approach, you can verify the OS patch before beginning the rebuild in step 8 and replicating it across the mirror.

---

NOTE If the logical drive fails and the server will not boot, you probably removed two drives that were mirrored to one another. Shut down the server if this occurs. Reinstall all drives and reboot. The drives mirrored to one another will be marked okay. All data should remain intact. Start again at step 1, being careful not to remove any drives mirrored to one another.

---

8. Insert the new drive(s) into the open drive bay. The array rebuild should start automatically. The replacement drives must be the same type (SATA or SAS) and the same or greater capacity as the original drives. If the rebuild does not start automatically, reboot and press F1 to enable rebuild at the 1786 Recovery Needed POST message prompt.

---

NOTE The drive-online LED will blink during the rebuild process. The rebuild is complete when the LED stops blinking and turns solid green. A 300-GB SAS (3.5” 15K) drive can take up to 2 hours to rebuild a mirrored array (depending on the number of IOPS the server is submitting to the controller).

---

9. Boot the OS after the rebuild completes. Rebuilding the array re-mirrors it and brings it back into a fault-tolerant state.

I'll see if I can find a newere version of the document

[SysAdmin2]

Thanks, will try if this work.

What is this isn't the case:

'1. Verify that the controller contains RAID 1 or RAID 1(+0) logical drives only'

F.ex. if we have a server w/ one boot raid1 and one or more other data non-raid1 set (5,6,50,60), could we then move one boot mirror side to a new server pick this of resync to fault tolerance and delete the other non-boot arrays [and possible create new data raid sets]?