Grooming of Exchange Sets does not remove old Fulls (by design?)

[jcapslock]

I am doing a full backup of the Exchange server once per week with differentials through the remainder of the week. I have the set configured to keep 14 backups. The differentials get groomed out properly, but the Fulls never go away, even after the last differential of the week that full was done in is groomed out. As a result I run out of space if I do not keep on top of it. Let me know if this is by design so I can write a batch file to delete full backups older than 2 weeks.

 

Mark

[Mayoff]

If you delete RDB data files from the backup set, your backups will be corrupted and you can't restore.

 

Differential backups don't work without a full backup. The differential is based off of the full backup.

 

If you feel safe, you can try to manually delete the full backups from Configure>Backup Sets>Backup Set Properties>Snapshots, but keep in mind the differential is useless without the full

 

I would recommend putting Exchange into it's own backup set and do a Recycle backup instead of grooming when you want to remove old data.

[jcapslock]

Mayoff,

 

Let me clarify. First I am not manually deleting MDB files, I am trying to groom them through the backup sets groom policy. I also am aware of the balance between full and differential backups.

 

I have our system set up to do full backups of the Exchange server once a week on Friday and take differentials for the rest of the week. My intention is to only have 2-3 weeks of Exchange backups online. So after 2 weeks I have 1 full and 6 differentials for each of the 2 weeks. Now on the 3rd Friday I take another full. So now I have 3 full backups. On Saturday grooming does nothing, as I would expect since the oldest full cannot yet be deleted because of the differentials that depend on it. The next week though, I would expect that the grooming operation would eliminate all the differentials from the oldest week (Now 3 weeks old) and the full from that week since all of the differentials that depend on it have been groomed out and it is no longer necessary. Instead the differentials do get groomed out, but the fulls just keep stacking up and never go away unless I manually do the operation you mention above. I had 5 full backups stacked up but only the latest 2 had differentials as I would expect.

 

Are my expectations just not matching up with the way Retrospect treats database grooming? I can do as you say, but it requires 3 times as many scripts and I just like to keep things as simple as possible.

 

Thanks,

 

Mark