How can we help you today? How can we help you today?

THREADCOUNT vs FILECOUNT

Hello All,

We currently use SQL Backup with the THREADCOUNT = 7 option

I have used multiple files in the past to speed up backups and am aware of the FILECOUNT option in SQL Backup

Still being fairly new to using SQL Backup any input from more knowledgeable people would be great.

How would this also impact our restore process using the GUI and also via SQL Jobs / Scripts for our verifaction processes?

Thanks in advance
BTreeToBTree
0

Comments

1 comment

Sort by
  • Eddie Davis

    Hi,

    The THREADCOUNT parameter specifies the number of threads the backup job will use when creating the backup file.  In background when SQL Backup issues the backup command to SQL Server, it creates the Virtual Device Interface (VDI) and waits for SQL Server to respond.  The number of VDI's created is dependent of the THREADCOUNT value.  This process actually creates a corresponding number of sqb files, which are then wrapped up into a single *.sqb file.  The THREADCOUNT value used to create the backup is stored in the *.sqb file header.

    The restore process will unwrap the single *sql file, back in multiple sqb files and generate the corresponding VDI's to perform the restore, matching the number of VDI's configured for the backup job.

    Using your example THREADCOUNT = 7, this will create backup job with 7 VDI created, then wrapped up into a single *.sqb file.  The restore process will create 7 VDI's to perform the restore.

    Valid integer values are between 2 and 32 inclusive.  However, you also need to be aware of the total number of CPU cores available.  As specifying a THREADCOUNT value higher than the total number of CPU cores will affect the performance of the backup task badly.

    The ideal value for the THREADCOUNT parameter is: (Total number of CPU Cores) - 1.  For example Total number of CPU Cores = 8, so 8 minus 1 = 7, set the THREADCOUNT parameter = 7.

    Unlike the backup task, the user cannot set the THREADCOUNT parameter for the restore job. The THREADCOUNT value is stored in the backup file header and the value stored here is used for the restore process.  If the total number of CPU cores of the server is less than the THREADCOUNT value, the performance of the restore job will be badly reduced.  So if restoring the backup file to a different server, ensure the hardware specification for the CPU of the restoring server at least matches the hardware specification for the server that created the backup.

    I hope the above answers your question.

    Many Thanks

    Eddie Davis

    Senior Product Support Engineer

    Redgate Software Limited

    Eddie Davis
    0

Add comment

Please sign in to leave a comment.