Hi all, Exchange 2010 mailbox DAG storage will on our HP Lefthand iSCSI (SAS drives) SAN but, generally speaking, is it better to have each database on it's own Windows volume or, because we're using a SAN, can we use one big Windows volume and put all the databases on it? We will be isolating the logs onto their own volume. Regards, Graham

Hi, yes of course you can, the best choice would depend on your scenarion. For instance in the future you might want to have a possibility (or need) to move a san volume through servers, so in that case you would have more flexibility if you keep each DBs on dedicated volume...Vincenzo MCTS, MCTIP Server 2008 | MCTS Exchange 2010 | WatchGuard Firewall Security Professional

Talk to your vendor about this. Generally there would be one database in one LUN and the logs will share the same LUN as each other but not with any databases. This all depends on what your backup programme will be though. "Birtybasset" wrote in message news:bd073271-2fec-485f-957a-ae3cc3ae976c... Hi all, Exchange 2010 mailbox DAG storage will on our HP Lefthand iSCSI (SAS drives) SAN but, generally speaking, is it better to have each database on it's own Windows volume or, because we're using a SAN, can we use one big Windows volume and put all the databases on it? We will be isolating the logs onto their own volume. Regards, Graham Mark Arnold, Exchange MVP.

Mark, In SAN terms, what do you mean by LUN? I'm no storage expert but when we ask the storage guys at work for a volume, we end up with raw Windows volumes in our VM which we then format as appropriate. Backup is done by HP Data Protector (VSS based).

Mark, In SAN terms, what do you mean by LUN? I'm no storage expert but when we ask the storage guys at work for a volume, we end up with raw Windows volumes in our VM which we then format as appropriate. Backup is done by HP Protector (VSS based).

1. Not sure how many DAG's you;re using, but it';s best to have each DAG in it;s on disk group for increased avaliabiltiy 2. I think the focus should be more on your your SAN setup rather than this specific question but will answer this too., i.e make sure that no other disk intensive app such as SQL is sharing your exch disk, make sure you have setup the best RAID as per your requirements and recommnendations from HP, probably recommend RAID 10. 3. I would contact HP for recommendations for your SAN setup. 4. See article below, will answer your questions and five you best practise on setup. http://technet.microsoft.com/en-us/library/ee832792.aspx#BestSukh

And as a recommendation, before deploying it into production, run Jetstress test on the storage so you will see that it passes the tests But start with contacting the vendorJonas Andersson | Microsoft Community Contributor Award 2011 | MCITP: EMA 2007/2010 | Blog: http://www.testlabs.se/blog | Follow me on twitter: jonand82

You could put all the database on one Windows Volume. But puting all the database on one Windows Volume makes the SAN a single point of failure. When designing your storage, you had better consider the points of failure, backup, and I/O. Please remember to click Mark as Answer on the post that helps you, and to click Unmark as Answer if a marked post does not actually answer your question. This can be beneficial to other community members reading the thread.

Well that's thing about the SAN, as expensive and clever as it is, we cannot avoid hosting all copies of our data on it. Well, we could host another database copy on separate DAS storage but we have no budget for that. Therefore at a minimum, we would like to create RAID0 volumes on each of the storage nodes in the SAN cluster. Both nodes are in geographically dispersed locations so at least we could separate the database copies instead of having them both striped across both nodes (network RAID). That would not only potentially increase performance but would be a more effective usage of SAN space. We intend to use both Jetstress and Loadgen before entering production.

Well that's the thing about the SAN, as expensive and clever as it is, we cannot avoid hosting all copies of our data on it. Well, we could host another database copy on separate DAS storage but we have no budget for that. Therefore at a minimum, we would like to create RAID0 volumes on each of the storage nodes in the SAN cluster. Both nodes are in geographically dispersed locations so at least we could separate the database copies instead of having them both striped across both nodes (network RAID). That would not only potentially increase performance but would be a more effective usage of SAN space. We intend to use both Jetstress and Loadgen before entering production.

A LUN is just the disk, whether the disk is a physical bit of tin in a server you can point to, something that your storage people present to you or the VMDK/VHD file that gets given to you. Its just the terminology for the thing that Disk Admin in Windows thinks is a disk. Sorry for the lack of clarity there. "Birtybasset" wrote in message news:bb9761f7-3f78-4651-aa5e-75266ebb6582... Mark, In SAN terms, what do you mean by LUN? I'm no storage expert but when we ask the storage guys at work for a volume, we end up with raw Windows volumes in our VM which we then format as appropriate. Backup is done by HP Data Protector (VSS based). Mark Arnold, Exchange MVP.