1

0

I’ve already implemented filestream on our exisiting table. However, I’ve a few questions to ask for FileStream implementation. Before that, here is the quick look of our environment. We’ve around 4-5TB of database which has around 5 BLOB tables. Rightnow, I’m focusing on only one table.The BLOB table size is around 320GB. The number of records would be: 758495. We’ve image size which is between 0kb to 19MB. Is it ok to implement filestream for this scenario? Does it imapact a performance in a long run?

I followed following steps to implement filestream in my test environment

  1. Enable FILESTREAM
  2. Create a FILESTREAM Filegroup
  3. Add a FILESTREAM File
  4. Modify the Existing Table - Add a FILESTREAM column
sp_RENAME 'fun.newspaper.Fdoc', 'Fdoc_Old' , 'COLUMN'`
GO
  1. ADD GUID columnyAlter table fun.newspaper ADD GUID UNIQUEIDENTIFIER ROWGUIDCOL NOT NULL UNIQUE DEFAULT NEWSEQUENTIALID()
  2. Migrate Existing BLOB Data Insert Existing BLOB Data into the FILESTREAM Column Check the FILESTREAM Data

In step 5, we’ve already have GUID column. Is it possible to update that GUID column instead of creating a new one. Here is my existing GUID column. In SQL, we couldn’t use modify instead of add in alter command so I’m thinking of right way to implement it:

Alter table fun.newspaper modify newsIdentifier...

[newsIdentifier] [uniqueidentifier] ROWGUIDCOL NOT NULL,

If so, could you please help me to modify existing GUID column instead of creating a new one.

Also, I would like to know when we backup the database. How to take exact backup and restore for filestream filegroup? In the future, if we need to migrate filestream, what is exact possible way to migrate it?

How to reclaim space after implementing a filestream?

If I would like to go with any training, could you please share a link for that.

2 Spice ups