We are using esxRangerPro ver 3.1.7.14927 in our production environment and are starting to test using 3.2.1.20699. We are using the Legacy mode for Ranger backups. For our run of the mill servers, we are getting reasonable compression and backup times. On database servers (SQL, Exchange 2007), the backups are greater than the original data size on disk; it's as if database backups are creating negative compression.
All servers are being backed up "hot". We are using a second backup solution specific to the databases, with Ranger having been a once a week "full" backup. The two backup methods occur never overlap. As a practice, we do not allow any of our VM's to have a snapshot for greater than 24 hours.
For the 3.1.7 version, we have tried turning on and off the quiesence switch without much of a change on the backup set size. Although today is my first pass at using the 3.2.1 version, turning off quiesence and invoking the VSS driver instead, it doesn't seem to make an impact on the backup size - the SQL server that I am testing with ballooned up to double the database size on disk and is still growing the backup size as I compose this.
That being said, is it expected that the backup size will be greater for Ranger backups of database servers than the original on disk data size? I haven't poked around on the restore side to see what it is that is generating the large backup files, thinking that it has to be how we are implementing the backups and not a "feature" of the esxRanger software. I also haven't been able to find any information that would indicate what type of compression to be expected from a backup of a database server that is being done "hot".
Thanks in advance for your patience and guidance.
Jim
Thanks in advance -
JimInMN