Backup Exec 2012 - Scheduling Dedupe Jobs

[jpk]

Hello everyone,

I'm new to using the dedupe option so I appologize if my questions are repetative..

How are you guys typically configuring your dedupe jobs so that you retain the data with fulls/dif/inc efficiently?  There is an option labeled "keep for", which I'm assuming is similar to the retention period on tape media... yes/no?

Is there an advantage to creating frequent full backups?  All I can figure is that decreasing the time between full backups would allow you to reclaim space if needed.  If I set a job to take a monthly Full with daily differentials, what should I set the "keep for" to?  For example, if I want to keep 4 months worth of history on a server, is this the correct setup:

Full - First Saturday of each Month --> Keep For = 4 Months
Differential - Daily --> Keep For = 4 Months

Thanks,

Josh

[SuperBrain]

Differential backups are such that when you want to restore something to a particular point in time, you need to have the last successfull full backup set AND the latest differential.

So, you can keep the Differentials for a couple of weeks because:
1. Differentials contain all data which was changed since the last full backup.

2. Differentials are larger in size as compared to incrementals.

So coming back to the "keep for":
 

Full = Keep for 4 months (as required)

Differentials = Keep for a couple of weeks should be ok (unless you want to keep them for 4 months)

FYI:

Keep for

 Designates the amount of time for which you want to keep the backup sets or job history.

See http://www.symantec.com/docs/HOWTO73365 for more info. 

[jpk]

Thanks.  I've read through the documentation and I have a good understanding of the differences between the job types.  My bigger question is what is a good plan for retaining the data within a dedupe folder environment.  Do most admins take a full backup and then run inc/dif jobs for a long period of time thereafter?  What's the advantage of doing multiple full backups within a relatively short period of time when everything is on-disk?  Isn't a Full + 30 days of incrementals/differentials essentially the same thing as a full everyday?

I'm trying to get a grip on how I should be processing my servers against the dedupe folder.  I have about 30 servers that i want to protect and schedule.  Unfortunately, I'm still battling issues on how slow things seem to be traversing the network.  Not to mention, system state backups are still brutally slow which kills my backup window.  I had high hopes that backing up to disk instead of tape would decrease my backup window but I'm not seeing much of that.

Unfortunately, I wish I had switched from BE to something else prior to purchasing the extra features/licensing a month or so ago.  I'm like many others on this forum in that a lot of features i used from BE2010 have been ripped from 2012.

[pkh]

 Isn't a Full + 30 days of incrementals/differentials essentially the same thing as a full everyday?

No  When you restore a full backup, you only need the full backup,  When you restore a full + 30 incremental backups, you need the full backup + the 30 incremental backups.  The time and effort for the second option is considerably more.

If your statement is true, then there would be no need to do any full backup. 

[SuperBrain]

For Full + 30 differential backups, you only need the FULL and the latest DIFFERENTIAL to restore the most recent information.

Do most admins take a full backup and then run inc/dif jobs for a long period of time thereafter?

Most Admins take :

A full backup either once a week or once in 2 weeks (but it still depends on lot of other factors

Incrementals/Differentials daily i.e. in between the full backup.

As far as retention period of the full backup is concerned, it again depends on factors like availability of storage/company policies/how critical the information is/etc.

What's the advantage of doing multiple full backups within a relatively short period of time when everything is on-disk?

Doing multiple full backups within short / reasonable period of time (days) helps reduce the restore time & effort. Also, it ensures good reliability on the backup sets (coz they are full backups)

Full + Incrementals = more restore time since every incremental set will need to be restored. Also if any one incremental set in the sequence goes corrupt, that breaks down the sequence and you will not be able to restore beyond that point

Full + Differentials = less restore time but, more disk space required. During restore only the latest Differential set is required along with the Full backup, hence, more data reliability.

Isn't a Full + 30 days of incrementals/differentials essentially the same thing as a full everyday?

No.

Full = Full 

Incrementals = Data that has changed since last Full/Incremental backup.

Differentials = Data that has changed since the last Full backup.

[Srineet_BE]

With the use of deduplication, you won't typically consume a lot of extra storage even with additional full (weekly, for example).

Also note that, Backup Exec won't expire a full backup as long as dependent incremental or differential backups are to be kept. So in your example, a full backup will be expired only after the differential of the last day of that month has finished its 4 month retention.