Question about the effects of defragmentation on Incremental Backups of files/folders

ATI doesn't rely of file time/date stamps to determine changes; it relies on sector level changes.

Defragging moves sectors (actually clusters of sectors). ATI will observe a changed sector location as a the file being changed and the entire file will therefore be picked up in an inc backup. Since defragging can move at least one cluster of a great many files, incs tend to be very big after a defrag.

Of course you can test this easily by defragging a single file that is other wise unchanged and then seeing if its picked up in an inc. The win defragger doesn't give you the option to defrag a single file but other defraggers do. Another way to test this is to change the date/time stamp on a file with something like the free program Attribute Changer ver 6. This will only change the file table entry without changing any file sector locations.

http://www.petges.lu/

Note that defragging generally is one of the most overrated operations one can perform on a drive. A drive has to be really badly (I mean really really badly) fragged for defragging to have any noticable benefits. So it's something you might want ot do only infrequently.

Thanks Scott for Your answer and explanations.

Hello Scott (and other experienced ATI users).

I wonder why it is on the basis of changed sector data, that ATI decides whether a file should be considered as having changed.

Other file backup Software are instead using the date/timestamp file attribute of the last change. Such a logic avoids to create large inc files after a defragmentation.

Personally, as a user, I would much prefer (for file/folder backups) a logic based on a date/timestamp compare...even, if the backup would take longer. For me, the amount of disk space used for the backup files is much more important than the time to perform the backup.

Probably, not all users have the same preference. It would therefore probably be nice, if the user could have a choice between the two methods.

Meanwhile, I have to decide whether I shall run (manually) defrag much less frequently (e.g. every 3 or 4 months) and perform after each defrag run again a Full Backup for all my tasks. Not very user-friendly...

I agree with Robert. Where is the documentation for not defragging every other day as with my Perfect Disk (Raxco)? Why would a reputable company seek to improve it's product every year if defragging often was unnecessary?

lol, well that brings up a few dozen or so reasons ;) . Perfdisk is a very good example. I use the product myself on all my machines so I am not disparaging the quality of Perfdisk. However, it's a very good example of a prog that comes out with a new for-pay version every year even if there is virtually nothing changed except for few doodads here and there.

As for the benefits of defragging, you can do the math on the amount of savings in average access time that one can appreciate by defragging a huge file that might be in hundreds or even thousands of fragments. Restores are done very infrequently, so saving, say, 500 msecs (or half second) on an hour long restore isn't likely to be worth the time it takes to defrag. Much more beneficial to defrag on database files that are accessed very frequently, system drives, etc. Even if you disregard the effect on ATI inc or diff file size, defragging is not a very good investment for image/backup files.

Raxco tells you defragging improves access time but they don't tell you how much time you save -- the amount is curiously small. If the small amount of time affects your file accesses many thousands of time a day, it can add up to something, but for those rare restores, the math just isn't there to support it.

STEVEN HOLTON wrote:

I agree with Robert. Where is the documentation for not defragging every other day as with my Perfect Disk (Raxco)? Why would a reputable company seek to improve it's product every year if defragging often was unnecessary?

I agree incremental backup should NOT consider the file location. Why user select incremental backup because it takes less time to do daily backup or frequently backup. I don't know whether my problem is caused by ATI considering the file location change instead of file timestamp. I put all my important data on an external USB hard drive, I carry the files between home and office. I always mount the USB drive as unique drive letter at home, I schedule ATI backup as incremental backup but ATI always do the full backup, it used up a lot of space and take a long time to complete the backup.

I didn't do any defragmentation on this external USB hard drive, but why ATI still consider all files are changed and do the full backup at last? Is it because of this hard drive is mounted from external?