Server for New Exchange Environment 2

[timmoat]

Hi all,

I was hoping that you would be able to offer abit of advice on a new Dell rack mount server for Exchange 2007.

Users: 65
Max Mailbox Size: 500mb Initially although once old GroupWise archiving is completed this will reduce to 200mb/250mb
Usage: Medium to Heavy usage

We are moving across from a Netware/GroupWise environment and it is logical to move to Server 2008/Exchange 2007. Despite the relativley small user base I think its best to have fileserver, dns, dhcp, domain controller, global catalog on another server leaving Exchange on its own box.

And why Dell? We have a Dell premium account and in my years as an admin I have never experienced any major issues. Swings and roundabouts choosing between Dell and HP from what I have read. PowerEdge 2950 seems a logical choice but I am no expert.

This is only to provide an initial idea and I intend on having a day of consultancy prior to purchasing everything to run over migration plan.

My greatest concern is over doing it of the spec side of things. I've come from an environment of 400 odd users so 65 is not a huge amount.

In terms of disk arrangement i'm thinking:

System RAID 1
Logs RAID 1
Mailboxes RAID 1/RAID 5/RAID 0+1?

Is the above over doing uit for just 65 users?

Would anyone have any recommendations for such a server config? 8 gig of RAM suffice?

Mailbox, Hub Transport, Client Access will be on the one server. No need for edge as firewall is locked down to only receive from our Mail Filtering provider. All archiving done on external mail filtering and Exchange 2007 journaling will provide archiving of internal mail upto email filtering provider.

Thanks for any feedback - if your interested in anything else please let me know.

ps for those migrating from GroupWise the Quest tools are a must! Working very well in test environment

 

[58sniper]

Put the mailboxes on RAID 1+0

No need for edge as firewall is locked down to only receive from our Mail Filtering provider.

That doesn't mean there is no need. Just less of a need.

Pat Richard MVP

Plan for performance, and capacity takes care of itself. Plan for capacity, and suffer poor performance.

http://www.ucblogs.net/blogs/exchange/

 

[timmoat]

That doesn't mean there is no need. Just less of a need.

Fair point. Again in a larger environment I would but 65 users doesnt really justify it when you take into account the firewall being locked down to only recieve port 25 from our email filtering ip range.

Have you got any other comments/feedback/advice regarding the above?

Is the above setup over the top for 65 users?

 

[SilviaS]

Have you considered LCR or CCR for resiliance
I am currently move a 5000 user base from GW 6.5 to Ex2k7
The migration tool works well but we have some pretty big user archives to move over to.

 

[timmoat]

Have you considered LCR or CCR for resiliance

Plan on using LCR, again as its the only viable resliance method for a userbase of this size with only 1 server in the mix.

With regards GroupWise I tormented myself for months working out the best way to do it. In the end ruled out co-existance as the connector/api just was not up to it and it would have caused more headaches for me during the migration period. Instead I am switching inbound mail across to new exchange server and backfilling mailboxes with 1 months data on migration weekend. Currently testing this... like crazy!

My users have lived in a world of no mailbox limitations and as such just 65 users has a GW post office of 80gig and an archived mail folder in excess of 100gig. Oh how things are going to change!

Will be using Quest to migrate everything apart from previous months emails to PST which will then be uploaded to 3rd part Archiving solution. Will then have a button in outlook to whisk users off to the archive portal in an attempt to pacify them.

Not as complicated as your migration but do you think the above sounds like a good approach?

 

[SilviaS]

We do have co-existance with no major issues, the quest migration tool works well, we then convert the archives to pst. Quest archive manager then pulls them into the application and hey presto its done.
If only doing was so easy....

Your way with 65 users over a weekend should be fine.
Good luck.

 

[timmoat]

Cheers!

Quick question: the Dell server I am looking at is the PowerEdge 2950 2U rack mount.

Now, it can only handle 8 disks if they are 2.5 inch...which means the maximum size for 15k SAS is 73gig. That should be fine for the logs and the OS but the mailstore could be an issue. To get 146gb I would have to use 10k SAS.

Would 10k on the mailstore cause any real issues? It will be a RAID10 array. Remembering that this is just 65 users.

Dell guy said it would be fine - but i'd like to hear from you guys that have experience outside in the real world.

Thanks again for any info!

 

[58sniper]

Should be fine. Transaction Log drives should be 15k if possible, but 10k store drives are no issue.

Pat Richard MVP
Plan for performance, and capacity takes care of itself. Plan for capacity, and suffer poor performance.

http://www.ucblogs.net/blogs/exchange/

 

[timmoat]

Should be fine. Transaction Log drives should be 15k if possible, but 10k store drives are no issue.

Thanks. Have changed quote to now include 15k drives for the logs.

Based on 65 users having mailbox sizes at a max of 500mb on RAID10 using x4 146gig disks should be fine i'm thinking.

Would you recommend RAID10 over RAID5 of rmailbox store? Is this due to performance gains?

A former collegue said RAID5 was easier to expand but in truth the server has a max of 8 disks and so I wouldnt be able to cram any in anyway!

 

[58sniper]

Yes, I would recommend RAID 1+0 over RAID 5. Plenty of responses here from XMSRE about RAID performance and Exchange.

Make sure you have enough space for a recover storage group.

Pat Richard MVP
Plan for performance, and capacity takes care of itself. Plan for capacity, and suffer poor performance.

http://www.ucblogs.net/blogs/exchange/

 

[Zelandakh]

Users x mailbox size = 65 x 500MB = 32.5GB. That's one store. Even at a 20% delta you are only looking at 6.5GB per day of TLs.

This is not going to tax a server of that size.

RAID10 is excellent for performance and resilience but poor at disk usage i.e. very wasteful. In your case it would be a good plan.

 

[Cstorms]

I used Jetstress to test my setup before a production install with a raid1 for OS, raid 1 15k for tls, and a raid10 for the dbs. I found almost a 100% gain in IOPS when comparing the two (i have 4 146gb 10k drives), I have 600+ users with a medium load on average and I have yet to see perfmon go anywhere near the 20ms threshold (actually havent seen a max of 10 yet either)

With a RAID5 you will get more space with your 4 drives but with a 500mb limit on 65 users you wouldnt even come close to the 270+GB you will have usable in the RAID10.. Easy choice

Good luck either way!

Cory

 

[xmsre]

Here's a reprint:

The rule of thimb is: If the write penalty of your chosen RAID type is higher than the read/write ratio of your application, then the RAID type is not appropriate.

RAID 5 has a write penalty of 4. Back in Exchange 5.5, when you had 5:1 or 6:1 read/write ratios with Outlook 98, RAID 5 was acceptable. In Exchange 2000, the read/write ratio decreased to 3:1 or thereabouts and RAID 5 was questionable. Some of the high end storage systems that cached the writes and reordered them, along with big cache on the storage controller could conceivably get to to 3:1 as long as you wre sticking to a small dataset, say a single 16GB database. In Exchange 2003, IO efficiency improved again, databases got larger, and you started seeing wide use of cached clients further lowering the read/write ratio. In Exchange 2007, it's not uncommon to see 1:1. RAID 5 just doesn't cut it any more. We're pushing the limits of RAID 10 (with it's write penalty of 2) in many Exchange 2007 deployments much like we pushed the limits of RAID 5 in the early days of Exchange 2000.

In the last 20 years, disk capacity has increased by several orders of magnitude. Rotational speed and random access times haven't improved anywhere near as much. My first hard drive, a whopping 10MB with 80ms access times was state of the art in 1985. Today, enterprise class SCSI drives hold 300GB or so and have 5ms random access times. Disks are getting bigger at a far higher rate than they are getting faster.

Spindle count for Exchange is primarily driven by the performance requirement. Determine the number of spindles using the performance requirement, then choose the size based on space requirement.