Daily differentials are doing a full backup instead

[nzamoeba]

I currently run a backup schedule with a daily on mon-thurs, and a weekly run on Friday.

Up until a couple weeks ago, we used to do full backups for both the daily and weekly, mostly because we could. Unfortunately the amount of data we're backing up now is starting to get pretty large so it was decided to switch to daily differentials, with one weekly full backup.

I've changed the Daily job to be 'Differential - Using archive bit', but still the jobs do a full backup of everything, taking just as long as they used to. Is there some setting I'm missing somewhere?

I'm using Backup Exec 11d for Windows Servers Rev. 7170

[Jani_P]

Hi nzamoeba,

I have a same issue, but I believe I know now why this is happening.

Nzamoeba, you did not state if your job was completed successfully or not.

You may have only one error, so that your backup job did not complete successfully. I'm having that issue. It seems that I cannot manage to get full backup complete 100% successfully (30+ servers, 1,9TB, 2 days backup time).  I did change as follows:

1) I have to run "Full-Using Mod.time" weekends and "Diff-Using Mod.time" weekdays because Linux doesn't support using archive bits. Linux file server is new to us.

2) I did merge my Windows, Netware and Linux backups to one job to minimize our Job policys and Selection Lists.

1)+2) -> After that change every weekdays diff. backup jobs are doing full instead. That is because one or two errors in backup job. Usually one or two files are "corrupt" on linux file server. They really are not, if I restore those from tape, they are fine (even zipped files open).

I did make a modification to registry so that job would complete successfully regardless of having corrupt files (only for testing, planning to go back to normal situation after corrupt files issue is resolved.

Anyway, that did NOT resolve weekly full backup issue, since BE itself encountered some RPC error and job did not complete success fully, although all the data was backed up. After that all the other servers are backed up full. I cannot let it run, because I do not have space for 5x full backup nor the time.

So the big question is (for Symantec tech's): If "Full-Using mod. time" backup job encounters only one failed resource (missing server, downed agent, one corrupt file, or BE internal error), why do you have to run full on weekdays on ALL servers?

If i have 99,9% of resources backed up, why does BE backup all those missing resources AND every other resource FULLY weekdays? Why it is not possible to backup fully only those failed resources? Should I do something differently?

For nzamoeba:

Page 239 in Backup Exec 12.5 Admin Guide:
"WhenBackup Exec runs a full or incremental backup, the time the backup launches is recorded in the Backup Exec database. The next time an incremental or differential backup launches, Backup Exec compares the file system time to the backup time recorded in the Backup Exec database. If the file system time is later than the database time, the file is backed up. When an incremental backup is run, a new time is recorded in the Backup Exec database. The database time is not updated with differential backups.

 ....

The Full - Using modified time backup method adds the time of the backup to the

Backup Exec database only if the full backup job completes successfully. If the

full backup job does not complete successfully, any subsequent backup jobs that

use the differential or incremental modified time backup methods back up all of

the data instead of just the data that changed since the last full backup. "

Kindly,

Jani P

[nzamoeba]

I'll try it with modified time to see if that solves the problem, too bad it'll take a week for me to find out.

Every job I've run has always 'completed with exceptions', and all those exceptions were due to 'file in use' issues. I've always ignored it because they were never terribly important (pid files etc) but I'd be pretty annoyed if that sort of thing prevented me from doing a proper differential backup.

[ltjimster]

Hello,

well why dont you try doing incremental backups instead, which backups up changes only and see how that goes.

hope this helps

Thanks
James

[Ken_Putnam]

Were you using the same selection list, or did the DIFF and FULL have their own lists?

[nzamoeba]

When I was doing full daily backups I had different selection lists, however I made sure to correct that when I tried to move to diffs and now they're both using the same list.

My weekly runs tonight so we'll see what happens after the weekend.

[nzamoeba · Accepted solution]

Changing to a differential using modified time has solved the problem, and it now appears to work as intended, with the diffs taking just under an hour instead of 10 for a full. I guess what we were doing wasn't compatible with the archive bit.