Since upgrading to 7.5 I have noticed my incrementals are very large. Policy=ms-windows, client=win2008r2...simple fileserver. The full ran in past couple days and completed normally. 1.5TB total data. I did some spot checking and saw that the archive bit was not reset. The client is set for incrementals based on archive bit. I manually removed the archive bit from ALL of the files. "attrib *.* /s -a". It removed the archive from all files. Now tonights Cumlulative incremental looks like it's doing a full. It is already up to 600gb of data. I just checked for archive bits again and there maybe approx. 30gb of changes since the full. I have many Large filservers at many different locations and am noticing this on everything I have upgraded to 7.5. Here is the interesting part, I run 2 backups everynight...don't ask. The backup to tape is the one not recognizing the cleared archive bit. The backup to dedupe disk is recognizing the cleared archive bit and just getting only the changes.
Checking with Windows Agent Engineering. There has been an intermittent issue with incrementals backing up more data than expected. Do you have TIR+ turned ON too...? What about Client Job Tracker...?
Also, let's keep in mind that if Archive bit is used for incrementals, do NOT run BOTH differential and cumulative incrementals...
We have seen similar issue from at least one other customer, but do not have a final solution yet. It is the opinion of Windows Agent Engineering to have you escalate your issue to Symantec Technical Support, since there is no quick answer here to resolve your issue through this forum.
You can though try taking incrementals based on time instead of archive bit, or try turning off client job tracker or accelerator, if you have that turned on, to see if these would help meanwhile.
There are only two configurations that by-design cause NetBackup incrementals to backup more data - they are time overlap and delta offset. Both of them have been ruled out as a cause of these NetBackup 7.5 Windows incrementals backing up more data than expected. So, this should be considered as a defect.
Also running 7.5, and seeing the same issue, but with VMware BLIB backups, using the new VMware policy. I am not sure that the cause is the new policy however.
Backup windows are now too small to finish in time, while there was more then enough time with 184.108.40.206.
Actually, the incrementals are pretty much the same size as a full backup now!
I am not seeing a specific pattern change on the Windows clients however!