Safe 'destroy' a RAID

Posted on 2011-03-15
Medium Priority
Last Modified: 2012-08-13
Hi folks.

I have a DELL server with an integrated PERC H700 and 2 disks in RAID1 (mirroring) for the Operating System and a SAN.

I need to test another operating system maintaining the existing one. I would like to:

- break the current raid
- place in safe one of the two hd's with the OS i want to maintain
- use the other hd to install a new os
- test
- rollback to previous OS using the original hd, recreate the RAID1.

Ok, I could make an image, store it in a safe place, reinstall the os then rollback using the image but is not easy to create an image with a raid configuration 'in the middle'.

The doubt is how to correct break the raid, and how the raid will be recreated (disk 1 over disk 0? vice versa? the coltroller asks for this?).

Question by:Lockwar
LVL 20

Expert Comment

ID: 35136893
The raid system is not designed for this.
From here on I am talking about general raid systems.  I do not know that model.
Once you break the raid array it looses the index / serial numbers for that setup.  The two hard disks can be handled seperatly.  You could wipe the disk and load the OS while keeping the old OS disk (removed) and safe.
When you come to rebuild - it normally will not allow you to do what you are trying to do.
There is a slim chance they have set it up to do this but I would work on backups and if it happens to work without this then you are safe.

Is this a working production server ?  If so you really should not be testing like this on it.
Get a simple pc and test on that.

Author Comment

ID: 35137010

This is not, right now, a production server but it will be... and i MUST use this server for testing purpose 'cause is already connected to san in a blade infrastructure. I can't use a simple pc/server for this purpose, sorry.

Is for that I asked for an expert :)

I know general raid system has troubles doing this job but I worked with other sysadmin doing this, long time ago.....

LVL 34

Expert Comment

ID: 35137224
The best and safest option would be to create an image of the system along with a backup(just incase), install the test os and when done simply reimage the system. There shouldn't be any problem creating an image even when using raid with a program like acronis. Another option would be to add a 3rd drive to install the test os on and just remove it when done.
Microsoft Certification Exam 74-409

Veeam® is happy to provide the Microsoft community with a study guide prepared by MVP and MCT, Orin Thomas. This guide will take you through each of the exam objectives, helping you to prepare for and pass the examination.

LVL 14

Accepted Solution

charlestasse earned 1000 total points
ID: 35137375
This is a vaild solution, pull out one of the 2 mirrored drives with the system down, removed the array configuration from the one still in the server at the controller, create a new Raid 0 on it and initialize the array. Then install and test your OS.

when you are done, boot back into the controller, delete the Raid 0 virtual disk, shutdown the server, pull this drive out and put your mirrored drive in, boot to the controller and verify that it see a degradded raid 1, boot to the OS and then add back the other drive, it will rebuild the mirror.
LVL 18

Expert Comment

ID: 35137556
You can always buy another drive off ebay or something and test with that, keeping your RAID intact.

Author Comment

ID: 35137600
For Jamietoner: I can't install a third drive 'cause I have on 2 bays..... and I need a software capable to create an image using a Dell PERC H700 controller... i'm googling around.

For Chariestasse: this is what I would do..... and I'll try this.. BUT after an OS image :) just in case.......


Author Comment

ID: 35137632
For chuckyh: yes, this is another solution.. but I would like to test the hardest one.... is not really a problem if I lost this configuration but is a good chance to test my skill and to learn something new in a cluste it environment......
LVL 47

Expert Comment

ID: 35137700
Your scheme is foolish, as you can not insure that you will neither pick up a bad block, or worse, a disk failure, resulting in partial or full data loss.  

You should really just buy another disk and use that.  Consider it as a necessity anyway, as you have no hot spare.

Author Comment

ID: 35138418
For dlethe: should be foolish on a production server, not really in a test server. Is not a problem, as mentioned above, to buy 1 or 10 new disks. This is 'cause I want to try this to perform a non standard task.
Not every day you can live in a 'perfect environment' and I think is better test while you're in a test environment... or not? :)
LVL 47

Expert Comment

ID: 35138682
I missed the part where you later wrote test server ...  If you were using virtual machines or software raid via LINUX md driver or solaris zfs, then this would be easy.  So if you are testing applications , then why not do this with virtual machines, then you don't have to blow anything away?
LVL 56

Assisted Solution

andyalder earned 1000 total points
ID: 35141832
HP have a document for a set procedure for doing this with their controllers, maybe Dell don't have it documented but it should be pretty straight forwards.

1. If there's an option to do an integrity check on the array then do it first, that'll ensure you're pretty safe from unrecoverable read errors.
2. Power off and remove disk 1.
3. Power on and it should boot, you may need to acknowledge the array is degraded.
4. To revert to other disk power off, remove disk 2 and fit disk 1
5. Power on and let it boot, put other disk in hot and it should rebuild (overwriting the disk that's plugged in when running)

Should you have problems booting you may have to go into the RAID BIOS and tell it to use the RAID settings stored on the disk.
LVL 10

Expert Comment

ID: 35144796
I use BESR (Backup Exec System Recovery) to create and restore images on a wide range of Dell servers with various RAID controllers.

I would be surprised if it couldn't cope with a H700.

Try the fully functional trial version first before you try anything drastic.

It could save you a lot of time and hassle.

I'd suggest making an image with BESR regardless of what method you decide to go with so you have a fallback plan.

Author Comment

ID: 35146297
For dlethe: The future use of one of this server will be an HyperV/VMWare cluster but now I must test some SQL Server application using all the 24 processors not virutalized.

For andyalder & Dextertronic: Thanks. I'll try both BESR and RAID test next week 'cause till Monday I'm out of office, thanks.

Author Closing Comment

ID: 35233224

Featured Post

Free Tool: IP Lookup

Get more info about an IP address or domain name, such as organization, abuse contacts and geolocation.

One of a set of tools we are providing to everyone as a way of saying thank you for being a part of the community.

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

New style of hardware planning for Microsoft Exchange server.
The main intent of this article is to make you aware of ‘Exchange fail to mount’ error, its effects, causes, and solution.
In this video, Percona Solutions Engineer Barrett Chambers discusses some of the basic syntax differences between MySQL and MongoDB. To learn more check out our webinar on MongoDB administration for MySQL DBA: https://www.percona.com/resources/we…
Whether it be Exchange Server Crash Issues, Dirty Shutdown Errors or Failed to mount error, Stellar Phoenix Mailbox Exchange Recovery has always got your back. With the help of its easy to understand user interface and 3 simple steps recovery proced…
Suggested Courses

850 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question