Exchange 2007 DB backup and restore

peter212693

peter212693

peter212693

Associate
Joined
28 Jan 2005
Posts
1,698
Location
Southport
Hi all,

We are planning a HDD upgrade on a clients Exchange 2007 server. Currently has 6x300GB drives in a RAID 1+0 and we are going to swap all 6 for 900GB drives. Now this will obviously kill the array.

Now the exchange is installed on the C: drive which is on its own array which wont be touched but he 3 databases and log files are on the array we will be swapping. Now we use Symantec system recovery as one backup solution taking a snapshot of the who system weekly (also use backup exec for daily tape backups). So I was thinking extra snap shot with system recovery of the array drop the new drives in, create the new array and restore the image of the database drive. Has anyone done a similar task and recommend a better way?

I was hoping to configure the new array in the server and then move all 3 databases from within the exchange console but there is not enough space in the server for all the new drives before removing all the 300GB's.

Cheers
 
Last edited:
Personally I'd run another Exchange server alongside, migrate everything to that, nuke the existing one and then migrate everything back once it's back online. You'd need an extra box to do it with but it keeps everything online whilst it's happening and removes the need to rely on backups. Hell, P2V it with an ESXi trial then install ESXi on their existing hardware and V2V back. Means it's no longer a physical box and makes future migrations a heck of a lot easier.
 
As above, opportunity to virtualise some of the setup CA/HT/ET's and stick the mailstores on a physical if you are worried about bare metal performance.

Who still installs software to C: these days and all on one server?
 
I did this a few years ago when I upgraded our drives from 300GB to 600GB. Mapped another server with enough space, moved the DB's within EMC, replaced the HD's then moved the DBs back. Obviously took a backup before hand but went smoothly.
 
I would definitely go down the route of migrating to temporary storage (using the EMC), replacing the disks and then migrating back. Obviously doing it all with backups in place.

Was never a fan of purposefully destroying a server and then restoring from backup, given that this is usually an untested path (what if your restore doesn't work).

It's hard to comment further without knowing what hardware you have available and how busy this Exchange server is. Although, generally speaking the Exchange DB's do not need to run on RAID10 :)
 
There is another way :

Obviously this is not without risk - you have been warned - but I have done it a few times
and make sure that the RAID controller support doing this!

Yank one drive - replace with a new one and let array rebuild
Repeat above until all drives replaced
Create new space on array
Expand drives at OS level
 
dmsims said:
There is another way :

Obviously this is not without risk - you have been warned - but I have done it a few times
and make sure that the RAID controller support doing this!

Yank one drive - replace with a new one and let array rebuild
Repeat above until all drives replaced
Create new space on array
Expand drives at OS level
Click to expand...
Yep, this definitely works (have done it on SmartArray a couple of times). But yes, pretty risky. It is a desperate measure.
 
How many drive bays does the server have in total, how many spares basically.

How big are the databases will factor into this as spare space is always going to be the limiting factor on what you can do. Not including the white space.

Id personally drop the extra drives in if there is space and create a new database, then migrate them over to the new database.

Otherwise I'd tend to agree that a second server would be the best option as working from a backup is never a good idea unless its tested regularly.
 
Last edited:
dmsims said:
There is another way :

Obviously this is not without risk - you have been warned - but I have done it a few times
and make sure that the RAID controller support doing this!

Yank one drive - replace with a new one and let array rebuild
Repeat above until all drives replaced
Create new space on array
Expand drives at OS level
Click to expand...

Done this many a time... Just check the hardware supports it and fail the disks within the software and shouldn't be a problem.

Or P2V if you have different hardware....
 
Thanks all,

All drive bays are full in the server so can add anymore without removing any. The controller is a P400 SmartArray controller and does support hot swapping the drives.

With the swap one drive at a time method. Once all the drives are swapped will the drive show un-allocated disk space on the drive which I can just expand to within windows? It does sound risky but too easy
 
Last edited:
Tried the swapping one drive at a time method today on one of our test servers out of curiosity and it worked an absolute dream.

Had 4 x 72GB drives in a RAID 1+0 running a 200 user DC. Swapped all 4 drives letting it rebuild between each and expanded the logical drive all withing 90 minutes!
 
Yeah... still risky as heck and talk about winging it. I presume you've got a very recent backup right before you go down that route right?
 
peter212693 said:
Yea we would have a backup from the night before. Not sure if we will be going down that route yet just wanted to test it out.
Click to expand...

I usually go with the safest route that doesn't require me to do a rebuild if it doesn't go well. Nothing worse than people asking what happened as if knowing the answer would help bring their system back online quicker. I'm taking the system offline so the upgrade will go well, deal with the short downtime! :p
 
You must log in or register to reply here.
Back
Top Bottom