Synthetic Backups

[mudabo]

I am trying to implement a synthetic backup on a client that over a TB in size, and has about 3 million files.   The  files do not have a high change rate, so this m,akes this client a good candidate for synthetic full backups.   I copied my production policy and cretaed to a new policy (and selected Collect True Image Restore Information with move detection)).   I made a one time full schedule, a daily incremental schedule, and a synthetic full schedule.   I ran the one time full backup and it lasted about as long as previous fulls.   However, when I run the incremental, it takes an extremely long time.   So much so, that it times out.   I ran an incremental  from my production policy, and it lasted 15 minutes.   Anyone have any ideas on  what why this    

[Yasuhisa_Ishika · Accepted solution]

True Image Restore(TIR) tracks which file has exist or not whereas non-TIR backup just tracks which file file is modified. This means TIR need to traverse all the files and dorectries even if no change has made to it. In most case, this behavior takes long time than traditional incremental backups.

Backup target with milions of files is not good candidate of TIR. Please consider to use Accelerator instead.