performance irregularities

[me123]

In testing, I backed up a test VM once every day over the last three days and found some weird results. The backup sizes of the three backups were: 29GB, 1.44GB, and 583MB which seems normal, however the throughput was about the same for all three backups and they all took about the same amount of time to backup. I cloned the machine and ran the defrag\sdelete script on these forums and re-ran the test, however the results were the same. Why is the backup of 29GB taking the same amount of time as the backup of 583MB? Should the incremental backups take less time and with better throughput?

I'm using the optimal compression and VCB SAN backup method.

[Gostev]

Currently, VCB is only capable of doing full backups. Incremental backup is additional functionality provided by Veeam Backup. To do this, Veeam Backup leverages VCB to pull the full VM from storage, find changed blocks on the fly (without data hitting disk on VCB proxy) by comparing the retrieved data to the previous state, and then save the incremental changes data. Since full VM image needs to be retrieved during this process, incremental backup takes comparable time with full backup (can't be faster than full VM retrieval time). Typically, on FC4 incremental backup is still faster (2-3 times), because you don't need to write as much data to the backup storage as during full backup (and backup storage is usually a bottleneck in case of FC4 VCB SAN backup). Hope this helps!