sbs2008 restore

rick81
rick81 used Ask the Experts™
on
i have  a sbs2008 server that needs to be restored. It has the disk setup below

Perc 6/i
2 discs RAID 1 - two partitions C:\OS, D:\Data
2 discs RAD 1 - 1 partition - E:\ contains exchange dbase, logs + users data.

The first RAID 1 has 2 disks with predicted failures.  So i am thinking ill need to replace both disks and do a bare metal restore from a backup.

Is the restore process as below.
1.  insert new drives, reconfigure raid 1 setup in bios
2.  boot from installation cd , repair computer, restore full backup
3.  load drivers for raid config, repartition drives and restore.
4.  Is there an option to restore specific drives to the new partitions?

i am thinking its best to delete both raid configs and restore the whole server.  rather than just the O/S partition as exchange dbase, logs are on the other drive.

are there any steps i am missing?  get the partitions the same and restore to each is my main concern, are there options for this is the restore.

thanks
Comment
Watch Question

Do more with

Expert Office
EXPERT OFFICE® is a registered trademark of EXPERTS EXCHANGE®
I would replace 1,let it rebuild,then replace the 2nd.

If this is a hotswap system,do it while the system is up and running,otherwise you will have to go through a disk import process.

As always ,make sure you have a KNOWN good working backup.

Author

Commented:
even though both discs are showing predicted failure.  DELL are telling me the drives have bad blocks and replacing 1 disk and rebuilding will most likely fail, if not just copy over the bad blocks.
I would test the backup to make sure it's good.

Generally speaking,disks by default have a spare pool of redirected blocks that get used whenever a bad block is happened upon.

When you get a predictive failure ,it's saying that you exceeded that threshold and will eventually fail.

Bad blocks do not get copied over,they get remapped.

http://en.wikipedia.org/wiki/Bad_sector
Should you be charging more for IT Services?

Do you wonder if your IT business is truly profitable or if you should raise your prices? Learn how to calculate your overhead burden using our free interactive tool and use it to determine the right price for your IT services. Start calculating Now!

Top Expert 2016

Commented:
Why have RAID if you are not going to use its capabilities? You've been spending double the time on writes for who knows how long, and getting faster reads and a halving of drive space and now you want to go through the anguish of a serve rebuild when all you need to do is a raid rebuild.

Author

Commented:
dont get me wrong i dont want to have to restore the server.  i am going on DELLs recommendations to replace both drives, rebuild the raid and restore from backup.

also various documentation i have read says you cant hot swap a drive and rebuild if the other drive has predicted failures.  both my raid 1 discs have predicted failures.  

is this not the case?
The worst that can happen is the rebuild will fail.

In which case, you follow Dells advice and reinstall from backup.

Author

Commented:
so back to my original question

Is the restore process as below.
1.  insert new drives, reconfigure raid 1 setup in bios
2.  boot from installation cd , repair computer, restore full backup
3.  load drivers for raid config, repartition drives and restore.
4.  Is there an option to restore specific drives to the new partitions?

i am thinking its best to delete both raid configs and restore the whole server.  rather than just the O/S partition as exchange dbase, logs are on the other drive.

are there any steps i am missing?  get the partitions the same and restore to each is my main concern, are there options for this is the restore.
If you are using the windows SBS backup utility ,you should not have to repartition drive.
As for the other raid set ,I'd leave it intact and not restore from scratch.

Author

Commented:
if the raid VD is recreated (when the 2 new discs are installed) though the existing partitions wont exist.  it would be 1 partition to windows, currently it has 2.

also my concern with restoring the O/S only is that the exchange dbases, logs, queues, etc are on the second raid 1 drive.  i thought it might cause issues as the OS backup will probably be 1 day or a at least a few hours earlier than the exchange dbases, etc.  would this cause any issues?
Top Expert 2016

Commented:
RAID has never replaced  backing up ever.  take your server offline, do the backups (don't forget to do a system state backup) , replace 1 disk from each raid array and then let them rebuild.   If the drives fail then you are no worse off.
Principal Consultant
Most Valuable Expert 2016
Top Expert 2014
Commented:
If the drives had already failed, then I'd agree that you need to do a restore.  But the fact of the matter is that these are PREDICTIVE failures -- meaning that they haven't failed yet.  So it is quite possible you can swap one of them out, and achieve a full recovery to the new drive and then swap the other out to then rebuild a non-failing RAID.

You should be aware that DELL PowerEdge Servers come with FREE LIFETIME Phone Support.  This little known fact can help you tremendously.  Because you can order the new drives and then get Dell support on the phone to remote in and THEY will hold your hand and even monitor the rebuild to ensure it is successful.

By attempting to do a RAID rebuild first, you avoid the problems that you are rightly concerned about regarding out of sync information.  It's actually less of a problem with your Exchange database being on a different partition, as opposed to workstations cached credentials which will sometimes be out of sync and cause the workstation to become deauthorized because they don't have an alternate DC to authenticate to.

I've seen Dell drives run for weeks with pending failures -- not to say you should allow this to continue -- but you definitely have time to attempt a RAID rebuild first, which is why they engineered this "Predictive Failure" warning to begin with.

If you do go about doing a Bare Metal Restore, just be sure you follow the official documentation so you don't leave out any important steps:
https://blogs.technet.com/b/sbs/archive/2010/06/11/how-to-perform-a-bare-metal-restore-on-small-business-server-2008.aspx

Jeff
First... You would not need to rebuild the OS. Make sure that you do build restore media CD that you will be able to boot from. DO test it.

I agree that re-building the Mirror by pulling one drive out and putting new drive in should work w/o any problems and then after it re-builds get the other old drive out and re-build again.

One difference though I suggest... Do the re-build off line meaning do not let the OS boot until the operation completes.

Author

Commented:
i wasnt going to rebuild the O/S.  worst case was going to restore from a backup.  boot off the sbs2008 disk 1, repair computer, restore from image, etc.

i'll try replacing 1 drive and hopefully it rebuilds and will go from there.

thanks for your replies, will let you know how it goes.

Author

Commented:
ok further to this question, i replaced 1 of the hdd's, installed a new one and let it rebuild the raid.  unfortunately it failed.  got to about 97% and crashed the server.  BSOD.

was able to force the other disk online and boot the server.  has been running off 1 disk, with a predicted failure for the last few days, risky i know.  looks like ill be installing 2 new discs, recreating the raid 1 config and restoring from backup.
I suggested doing the re- build off line. (Before the OS boots).
Is there a reason for not doing so? You may me able to complete the rebuild 100%...

Author

Commented:
We had a dell technician go onsite and do the rebuild. Dell tell me even if the disk successfully rebuilt the issue would come back due to the predicted failure errors on both disks.

Do more with

Expert Office
Submit tech questions to Ask the Experts™ at any time to receive solutions, advice, and new ideas from leading industry professionals.

Start 7-Day Free Trial