Exchange Storage Problem

Pigeon_Killer

Pigeon_Killer

Pigeon_Killer

Caporegime
Joined
21 Nov 2005
Posts
41,328
Location
Cornwall
Won't bore you with all the details so I'll keep it simple. My Exchange server has just about run out of space and I need to install two new 600GB drives in a server with only 1 working backplane.

1 x HP Proliant DL360 G6 Server
2 x 300GB SAS drives in RAID1 for System
2 x 300GB SAS drives in RAID1 for Data
1 x 280GB Exchange DB

My plan is to create an iSCSI target on another server which has 3 x 450GB SAS drives in RAID5, move the DB and public folders using EMC, install the 2 new 600GB drives and move the DB back.

I've also considering installing the drives in an identical server, moving the DB then installing the drives in my Exchange server and importing them as a foreign RAID set. Never done this before so not sure how well it works.

Anyone got any better ideas?

Virtualisation and SAN in next years budget btw! :rolleyes:

Thanks :)
 
Could you...

Backup
Backup again
Pull one of the 300GBs in the data array, replace with a 600GB
Let the array rebuild
Pull the other 300GB, replace with another 600GB
Let the array rebuild
Expand the logical drive (does the Smart Array have BBWC?)
Extend the volume in Windows (which version?)
 
Thanks for the suggestion. The Smart Array does indeed have BBWC but HP's user guide suggests it can take up to 15 minutes per gigabyte to rebuild and I need to do this over a weekend so it put me off going down that route.
 
A long time :p

I ran a test a few weeks ago and 4GB didn't take more than a few minutes. Also tried copying the same file to a SATA disk :eek:

Edit: I suppose it wouldn't matter when I swapped the drives using your method because the server would remain up. Just wouldn't have a job if the good disk died before the array was rebuilt.

Edit 2: Just realised 15 minutes is the non cached estimate. It's around 30 seconds per GB with cache :cool:
 
Last edited:
Decided to try my iSCSI plan and swap the drives one by one if plan A went wrong. Started at 9am this morning, took 5 hours to move 270GB, 1 hour to change the drives and create the array and 45 minutes to move it back.

Now have 253GB to play with \o/
 
Last edited:
Eulogy said:
Good stuff, you'd better set up some monitoring so you don't get blindsided again though!
Click to expand...
Been monitoring and managing it for nearly 6 months unfortunately. 80% of my users have a mailbox smaller than 1GB and delete/archive old emails when asked to but the rest don't really care. Just reached a point last week where I thought it would be easier to upgrade the disks than waste anymore time trying to manage them.

The next battle will come when we get our SAN and look to create separate storage groups with limitations. That will not go down well at all :o
 
Creat a maximum mailbox size, set it to email at x space then stop them able to send or recieve at y space.

Users don't appreciate how much space they take up.
 
Exchange 2007 I assume?

If you got more than 100GB of mailbox data you should be splitting out into separate SG's/DB's anyway. Best config is one DB per SG, even for PF databases if you have enough free SG's. Limit of 5 SG's for Exchange Std edition, 50 for Enterprise.

Max recommended DB size on Exchange 2007 is 100GB, for a CCR cluster it's 200GB.
 
Ah you mentioned SG's which are specific to 2007 as 2010 only has databases. But I know that in SAN's you can also get Storage Groups/Raid Groups etc so guess you may have meant that! :D
 
You must log in or register to reply here.
Back
Top Bottom