SAN storage space - how does it work

[richw1234]

I have been given the job to find a SAN solution for a customer, previously we have only had to deal with space upto 2TB but now we are looking at 20TB! I have seen a few HP solutions which fall into this category but I am not sure how the storage space is configured. A HP MSA1000 can hold upto 48TB but can I have this as one partition? If not, can the partitions be reliable glued together to make one big partition?

How any help or pointers would be much appreciated.

Richard

 

[pshedges]

There is a limit for a maximum LUN size of 2TB,this is a limitation of the controller this can be 'Pooled' together on the host via 3rd party software or through the OS's Volume Manager.

Different OS's have different limitations on what LUN sizes they can 'see'.

Below is a link which has information on Windows and how it can now support up to 256TB LUN's

http://www.microsoft.com/whdc/device/storage/LUN_SP1.mspx

But there is still a limit within the controller firmware on HP EVA's/MSA's for example of a maximum LUN size of 2047 GB (2TB), so you would need to create multiple LUN's on the SAN and have the host OS manage the pooling of them into one LUN.

 

[richw1234]

Thanks for your post,much appreciated.

Does anyone have any experience in glueing volumes together with Windows or HP software? Is safe how reliable is it?

 

[comtec17]

Concatenating drives is not a good practice. It works great for performance but sucks in a disaster or a lost disk.

You should make your LUNs large enough to house your data, even though not always possible, it is the best approach.

 

[richw1234]

Surely if I am Concatenating volumes, a failed drive should not matter because of the raid/spare scenario. I am not convinced a failed drive should cause any issues but I am not aware of any other pitfalls that there might be with this activity.

 

[pshedges]

You are not 'Cocenating' physical drives but 'Pooling' Vdisks into a single LUN on the host.

Drive failures will be managed at the SAN level, and with the RAID levels, disk group protection etc, drive failures should be transparent to the host OS.