On Tue, 24 Jul 2007, briandiven AT northwesternmutual DOT com wrote:

> For those of you that are seriously interested, here is the actual format 
> taking advantage of policy and schedule names that made our life easier.  I 
> should also state that we stood up extremely well to 5 audits over the past 4 
> years (BCP/vaulting audit, internal audit regarding records retention, backup 
> audit, internal SOX, and external SOX audit).
>
> Policy name example:
>
> Sybase-alderaan-PDS_SY24-model-DB ... Which tells me this is a sybase DB on 
> physical host alderaan on database server PDS_SY24 for the model database 
> instance and that this policy is a DB backup (vs. a log).
>
> Our audit requirements are for 30 and 90 day retentions and we send all 
> databases less than 25 GB to a D2D pool.  To accomplish this, we use the 
> schedule name.
>
> Schedule name example (There are 2 automatic backup schedules and 4 
> application backup schedules per policy):
>
> Automatic Backup Name:  PDS_SY24+model+30day+DB+tape+1 and 
> PDS_SY24+model+90day+DB+tape+1 ... Which tells me database/instance, the 
> retention, that it's a DB backup, destined for tape with 1 stripe.
>
> The key here is that we have a single script to maintain for the whole 
> environment, because it has all of the information to parse.  The DB team is 
> required to keep a table of all databases and whether they are active or not 
> and how big they are.  We activate/deactive/create policies based on their 
> table and the script determines whether they should go to disk or tape based 
> on the size.
>
> Application Backup Name:  There are 4 of them, 30day-tape, 30day-disk, 
> 90day-tape, and 90day-disk.
>
> I would also add that rerunning failed backups is one thing, but what about a 
> backup that never runs?  It doesn't show up on a failed rerun script.  Part 
> of the summary reports show databases that haven't had a backup in "X" number 
> of days so we catch those too.  Now the onus of the audit is on the database 
> teams to keep their table current and it is a very well documented, specific, 
> and verifiable process.  I wrote my own SLA's at a 95% backup success rate 
> and 100% restore success rate and haven't missed them for 2 years now.

How often do you perform restores?  What types of tape medium do you
use?  What robots are in use?

I find 100% restoration rate very nice; however, how do you achieve
that, I assume you have two copies of most pieces of data as mentioned
above 30/90 days?

Justin.


_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu