Restoring a SQL Server Database – A Practical Example
Performing a Database Restore is a part of a DBA’s daily life. A DBA may need to perform a Restore due to various reasons such as recovery, refreshing a database for testing purpose etc. Many times it can be difficult to perform a Restore due to corrupted media, low disk space on the server and so on. In this article, I will outline one approach which I used to Restore the backup of the Production Database a test environment.
A couple of days back, members of my support team approached me saying that they are unable to refresh a database named ABC on the OLTP development environment with the copy of the backup of the same database from the Production server. The backup copy taken from the Production server was around 75 GB in size. On our dev machine we were had just 1 dedicated drive(D) for SQL Server which was having just 70 GB of free space left on it.
After having a detailed look, I came to a conclusion that I could not free up any space on the D drive of the dev machine. One important point to mention is that our entire dev machines were in a different domain than the staging and production boxes.
I logged on to the Production Server and decided to split the backup of the database named ABC into two equal parts using the T-SQL as shown below:
BACKUP DATABASE ABC TO DISK='B:\DB Backups\ABC_1.bak', DISK='B:\DB Backups\ABC_2.bak' GO
Once the above block of T-SQL statement is executed, it splits the Full Backup of the database named ABC present on the Production server into two equal parts. For example, if the size of a database is 76 GB then it will be divided into two equal parts each of size 36 GB.
Once the full backup gets split into two equal parts, I can then perform the RAR on them. Obviously first performing the RAR and then moving them to the different server would be much faster than moving the original copy of a much bigger size. Once the split files were zipped successfully I then moved them onto my staging server. This is because the staging box was pretty good in terms of resources and also since both the Production and Staging servers were in different Data Centers, due to good hardware the copying process worked much faster. As discussed earlier the free space available on the dev OLTP box was just 70 GB whereas the backup copy was 75 GB therefore it was not possible to transfer the zipped copies of the full backup to the dev OLTP box. I was having another box which was used as a SSIS Dev server with large amounts of free space. They are as follows:
Drive D had 49.9 GB free space
Drive C had 55 GB free space
The dev machines were standalone machines and not a cluster, therefore there were no issues for in copying 1 zipped file of the backup onto the C drive.