Here I go again...drive config help 4

[Davetoo]

Building another Exchange 2007 server..going through the spreadsheet to determine best drive config and getting some odd results. So...I'd like some real world scenarios from you if I could please.

Mailbox count is 450 with a 10% growth over 5 years. Usage per mailbox is fairly minimal...25 in/out max per day as an overall average.

I have an IBM 3550 with 3.5" drives using two 146GB RAID 1 for the OS. Going to use the other four slots for my RSG later on.

Connected to a DS3200 with 12 300GB drives.

IF you were going to configure the drives, how would you do it?

My consultant is saying build 11 RAID 5 with a HS...I think that's a huge mistake, thus my query here.

I'm Certifiable, not cert-ified.
It just means my answers are from experience, not a book.

There are no more PDC's! There are DC's with FSMO roles!

 

[Davetoo]

The expansion box arrived yesterday and I did some moving around of the drives and reconfiged a bit.

My final config is:

Stores:
Spindle Config: 8 x 300GB, RAID 10
Partition Config: 7 x 159GB

TLs:
Spindle Config: 4 x 146GB, RAID 10
Partition Config: 7 x 38GB

Now, I had purchased some more 300GB drives for the RSG, however, it struck me that using them in a RAID was a waste of drives. The only time you use the RSG is when you're recovering data...which means the data is already stored/backed up somewhere else, so if the drive fails during a recovery it's not an issue. So, I'm going to use one 300GB drive for my RSG drive.

I'm going to then use 4 x 300GB, RAID 10, for an LCR of several of my databases that contain some high profile mailboxes.

I still have seven slots open on my expansion chassis for future expansion of databses/TL's.

Just wanted to close the loop on this. Appreciate everyone's help.

I'm Certifiable, not cert-ified.
It just means my answers are from experience, not a book.

There are no more PDC's! There are DC's with FSMO roles!

 

[Zelandakh]

That looks like a good solution although you'll be using the store spindles concurrently when Exchange accesses 2 stores at the same time. Nonetheless you'll probably never notice that.

Given what you've got to play with, that looks like an elegant solution.

Microsoft would no doubt have one answer but MVPs also live in the real world and I've seen far worse.

 

[Davetoo]

I just couldn't justify 28 drives for my stores (7x4 RAID 10). I thought this to be a good compromise and with the equipment I have in place if, in the future, I need to move stores to their own spindles, then I can quite easily and without disruption to my users.

Life in the real world huh? lol

I'm Certifiable, not cert-ified.
It just means my answers are from experience, not a book.

There are no more PDC's! There are DC's with FSMO roles!

 

[Zelandakh]

In the real world we can't have RAID10 for each store and another one for the TLs, 1 for the RSG and 1 for the little boy who lived down the lane...

Save the money from the other array and get yourself a big bonus instead

 

[xmsre]

For a random workload, as long as you size appropriately comingling doesn't have a lot of impact. For serial workloads, your IOPS throughput can drop as much as 6X when you comingle. When you look at storage management complexity vs. spindle count, I usually end up with one pool of disk for the DBs and one pool of disk for the logs. just use the random IO when running the performance numbers for sizing the log spindles. In Exchange 2007, Log IO is about equal to DB writes (about half the IOPS to the DB if you're using cached mode clients). The storage I work with day in and day out has no write penalty. For RAID 10, you'll need to factor a write penalty of 2 into your decision.

 

[Zelandakh]

The storage I work with day in and day out has no write penalty"...

I haven't got one of those but Christmas is coming and if you are stuck for ideas for me...!

I agree with xmsre.