MSA1000 Configuration Best Practices

[Neil Thirsk]

We have an MSA 1000 with redundant controllers and switches and 3 additional shelves of 146GB disks.
Multiple hosts will be configured on this SAN (Exchange and Vmware ESX)

Is it best practice to build the LUNS in the similar fashion to an EVA starting at the top shelf and working down that column then moving to the top of the next column of disks and so forth or is there a better way ?

Also what about array configuration ? Is it a good idea to have separate arrays for each host that will be on the fabric ? Or have a single array for all hosts baring in mind that both ESX and Windows 2003 will be on this SAN ??

Lastly we noticed that once a logical drive is created below another you can't delete the above logical drive without deleting the one below it, is this by design or am I missing something ??

Thanks in advance for the assistance.

[Khairy]

hi neil,

msa1000 is diff compare with its sister EVA. MSA1000 utilize normal scsi disk while eva is totally fibre channel storage.

So, to get better performance, create lun across multiple bus.

Msa1000 already comes with built-in dual bus disk backplane.
MSA1000 could support up to 4 scsi buses. So if you want to add more disk shelf, max is 2 single-bus msa30 or 1 dual bus msa30. You have to choose either one.

For scability, you could add 2 single bus msa30 which could give you additional of 28 disks. Overall with the built-in msa1000 disk backplane, you could have a total of 52 disks.

For better performance, you could add 1 dual bus MSA30 only but this will limit the msa1000 to have a total of 28 disks only.

Based on my experience and other poster in ITRC about msa1000, it would be good if you could separate arrays for each host in the fabric.

If you really want the arrays to be share with other host, try to limit number of host accessing the arrays.

Its only my 1 cents opinion.....

Rgds