'Preprocess' of CBT possible?

[Steve_za]

Case ID # 0544795
Is it possible to 'preprocess' CBT so it doesn't run during the backup? The initial CBT takes so long that it makes the backup run for far longer than usual, and then the snapshot is huge so it takes hours to remove it. It would be ideal if the initial CBT could be 'preprocessed' so it wouldn't have to keep the snapshot open while it did it.
Is this possible?
To give you an idea the VM is 3.6TB, I'm adding another 3TB which will need to be read for CBT on the initial run. Currently the backup takes about 5 hours, and removing the snapshot also takes about 5 hours, as it is a massively changing MS SQL server. Adding another 3 TB, even when blank will make the backup run for hours longer and the snapshot removal will run well into production time., which causes massive backlogs on our cacheing servers.

Thanks
Steve

[foggy]

Steve, it's not CBT that takes long, it's actual read of the 3TB of data you've added (CBT just returns the blocks that required to be read). This massive amount of changes should be read and transferred to the target and definitely requires considerable time.

[Vitaliy S.]

The initial CBT takes so long that it makes the backup run for far longer than usual, and then the snapshot is huge so it takes hours to remove it. It would be ideal if the initial CBT could be 'preprocessed' so it wouldn't have to keep the snapshot open while it did it.

Change tracking is used when you create a snapshot during the backup job, so the initial full run after you add new disk will take time to read the entire disk anyway. Not sure I fully understand what you mean by saying "preprocess"?