Alerts if disk is over 90%?

"However, when you delete something, there is no standard way to tell the array those blocks are now unallocated."

Hopefully, SCSI unmap will make its way from vSphere to other popular x86 operating systems.

I guess the only solution I have is to set my Exchange 2007 databases (16 DB's) to thin provisioned, it is a PS5000 with SAS disks (sorry not a PS6000).  The thing is our consultant set them to thick when setting up a couple of years ago, will I have issues with setting them to thin?

Thanks Don that makes perfect sense to me.  Our Exchange Databases never go down in size anyway (wish they did).

The other issue I have is the consultant who set up our Exchange 2007 server complately over sized all the database volumes.  For example a 20GB sits in a 100GB volume and another 8GB database sits in a volume of 100GB etc and these databse are 2 years old, so you can see the speed of their growth.  So I've been wondering how I can get that space back as I can't reduce the size easily it seems from the GUI, so I thought if I thin provision the volumes I wll get the space back, but I read somewhere that you might lose performance when the volume needs to ask for that extra space each to it writes to the volume, true?

To reduce the disks I did read that you can telnet onto the Equallogic and reduce the volume size and then do the same on the WIndows 2003 Exchange server using diskpart (I think), but this sounds scary!

It would be great to hear your thoughts.

Thanks  

Great,  Are there any performance differences between thin or thick, our SQL guy seems to think so. I have to set his disks to thick all the time, but to be honest I've not seen any speed degrade with thin provisioned disks.

Thanks, I will put the shrink to one side I think.

However just setting the thick volumes to thin interests me. Are you saying I should snapshot the volume straight away before converting and I should set it to 30%?

Also how would I recover using the snapshot should it fail?