B2D Directory Structure

[networkn]

Hi There!

Could someone assist me with understanding the directory structure of the B2D directory?

I have a lot of BIG (4GB) BKF Files which I presume make up the full backups, and then a bunch of directories called IMGxxxxxx directories which significantly vary in size from about 13GB to 105GB Interestingly enough they have increased in size from a certain date forward, but we have not increased our data significantly in that same timeframe.

Which files can I remove and what happens if they are part of a previous but still active set? Will they just recreate?

[pkh]

Normal backup use .bkf files and GRT backups use .img files.  You should not just delete them.  If you have set up your OPP correctly.  They would be overwritten correctly.

Do make sure that you go to Tools ---> Options and set the followind

[Colin_Weaver]

BKF files contain standard remote agent backup sets of file system and database data.

IMG folders contain GRT enabled backup sets of Virtual Machine and Database data

[networkn]

Ok well the problem is that my disk keeps getting full. I already have the option set as suggested by PKH. I think it looks like because we have Windows Deployment Services, incremental backups are 100+GB a day, because it's backing up the whole image directory. I am deciding if I should just delete or find a way to clean out the entire B2D directory and see how many a day are occuring.

[Colin_Weaver]

As well as the option pkh provided, also make sure any GRT enabled jobs are set, in the job/template properties, to overwrite media at the start (making sure that your media set overwrite protection is long enough for your data protection requirements)

In some cases if you do not set overwrite at the start of the backup job, but have it set to append, then IMG folders that have gone beyond their overwrite protection period will not be deleted from the system which can cause disk space problems.

Note that this is a BE2010 R3 and earlier comment, BE 2012 is fixed to overwrite at start for disk based jobs so should not have that problem.

[networkn]

Hi There!

I run BE2012 so this should be working properly. 

If I change the overwrite protection from 3 weeks to 2 weeks, the very next time it runs the job (at the start) it should go through and remove any backups (Inlcuding GRT) older than 2 weeks old, is that correct?

The only thing I have compatible with GRT is SQL, I don't run exchange or sharepoint, and I am only ever going to want to restore the whole DB so should I turn off GRT? If so do this, next job that runs, will it clean out the GRT Folders so I will recover more space?

[Colin_Weaver]

if you turn off all forms of GRT it won't clean out any IMG folders (ever)

Basically the IMG reclaim (deletion) occurs at the start of an GRT backup set as a specific operation/process. A non-GRT job will never trigger this reclaim process.

One other point however if you turn off GRT all you will get is an equivalently sized BKF file to hold the same data (although) compression may affect the actual space used by the bkf.

[networkn]

I don't really understand I guess. If I want to recover the space the GRT is using (not reuse) then what do I do? Doesn't seem like that would save me any space though.

Anyone know how to stop BE Capturing the WDS in incrementals, seems to be part of system state.

[Colin_Weaver]

OK I did a little more digging - up to and including 2010 R3 we woudl not have remoevd IMG folders if you stop running GRT jobs, from 20120 onwards the DLM (Data Lifecycle Managerment) process would groom out the IMG sets once they become overwritable, although DLM will not delete the last copy of a recovery point chain so some IMG folders might remain if their content meets that particular rule.