S
Someone
Hi all.
We're currently testing a new drive array. We're seeing some strange
problems with NTFS volumes mounted as folders instead of as drives. Our
current infrastructure with our current drive array has us creating a LUN
off of the array and mounting the LUN as a folder named after the customer.
Example: E:\customers holds all of our customer data
Mount LUN1 for CustomerX as E:\customers\CustomerX. The E: drive is about
12Gb, while the CustomerX LUN is 100Gb.
Once the LUN has been mounted, we would copy over the SQL data files and
attempt to mount the SQL database. The database itself is larger than 12Gb.
This is where it gets strange. With our current disk array (EMC Clariion
FC4700), everything goes normally, the database comes up and we start
processing.
With the array we are testing with (Chaparral RIVA), SQL aborts when we
attempt to mount the data files and says that the target doesn't have enough
space. It appears that SQL is looking at the size of the host drive only,
instead of the LUN or the sum of the host drive and the LUN. The only way to
get around this is to map a drive letter to the LUN. This paradigm is not
supported in our current infrastructure, so is not a valid workaround for
us.
Does anyone have an idea of what is going on here and how to fix it?
We're currently testing a new drive array. We're seeing some strange
problems with NTFS volumes mounted as folders instead of as drives. Our
current infrastructure with our current drive array has us creating a LUN
off of the array and mounting the LUN as a folder named after the customer.
Example: E:\customers holds all of our customer data
Mount LUN1 for CustomerX as E:\customers\CustomerX. The E: drive is about
12Gb, while the CustomerX LUN is 100Gb.
Once the LUN has been mounted, we would copy over the SQL data files and
attempt to mount the SQL database. The database itself is larger than 12Gb.
This is where it gets strange. With our current disk array (EMC Clariion
FC4700), everything goes normally, the database comes up and we start
processing.
With the array we are testing with (Chaparral RIVA), SQL aborts when we
attempt to mount the data files and says that the target doesn't have enough
space. It appears that SQL is looking at the size of the host drive only,
instead of the LUN or the sum of the host drive and the LUN. The only way to
get around this is to map a drive letter to the LUN. This paradigm is not
supported in our current infrastructure, so is not a valid workaround for
us.
Does anyone have an idea of what is going on here and how to fix it?