Implementing B2D2T, need advice on tape management

[steve2012]

Have BEWS 2012R3 installed and working.  Previous tech/admin manually created daily and weekly B2D2T jobs (Backup to disk with triggered copy to tape).  Now I need to create new Monthly jobs.  Problem is there are over 30 unique jobs and I'd rather not create 30 new jobs plus have to think about the tape management, offsetting start times, etc, so I was hoping to create a single new policy job that does the following:

Custom B2D2T Policy Job:

• Monthly B2D (full, append, 1 week retention, B2D pool 1, 1st saturday of each month)
• Weekly B2D (full, append, 1 week retention, B2D pool 1, 2nd to 5th saturday of each month)
• Daily B2D (incremental, append, 1 week retention, B2D pool 1, Monday-Friday)
• Copy Monthly D2T (linked to Monthly B2D job, append to tape, writes to Monthly tape pool, 11 month retention)
• Copy Weekly D2T (linked to Daily B2D job, append to tape, writes to Weekly tape pool, 4 week retention)
• Copy Daily D2T (linked to Daily B2D job, append to tape, writes to Daily tape pool, 1 week retention)

Note: by retention I mean using OPP (overwrite protection policy) to protect the tape or B2D file from being overwritten for the desired length of time.

What I cannot figure out is whether it is possible to configure the policy such that the first job scheduled OVERWRITES the first available tape and APPENDS the remaining jobs.  Currently the jobs will fit on a single tape (LTO-5 is ~1,5TB), so I'd like to avoid spanning media if possible.  If this cannot be done at the policy level I guess I could create a "starter" job that overwrites the first tape and the policy jobs append to it (media selection logic seems to choose the last used tape if it has room available).

More info:

• B2D pool is two B2D targets permitting up to 4 concurrent executions (example: 4 B2D jobs or 3 B2D jobs plus one B2T job)
• B2D storage capacity is a little over double the size of the Full backup, but not wnough for two full copies plus incrementals
• Single drive tape library (LTO-5)

On a related note, if using a policy to generate jobs, does the last job have to complete within 24hours of the first job starting to avoid timouts (execution window)?

For example: I have 30 jobs, the first starts at 6pm Saturday, the last job starts at 5pm Sunday (has to wait for the tape drive to be available after the first 29 jobs complete), would it have to complete the backup to disk and copy to tape within 1 hour, or does the backup window get adjusted for each job based on their start time?

[pkh]

What I cannot figure out is whether it is possible to configure the policy such that the first job scheduled OVERWRITES the first available tape and APPENDS the remaining jobs.

If you are using different policies, then you just need to change the template in one of the policies to overwrite and the rest will specify append.  If you are using just one policy to generate your jobs, then you would not be able to change the overwrite/append setting when you generate the jobs.

does the last job have to complete within 24hours of the first job starting to avoid timouts (execution window)?

No.  The execution windows is the time the job got to wait before it is started.  If all your jobs starts at the same time, say 10 p.m., then your last job must start by 10 p.m. the next day.  Otherwise it will be canceled.  If it starts at 9.59 p.m. on the next day, it will contimue until completion.

[steve2012]

Hmm...will look into creating the second policy (so obvious now that it has been pointed out).

My past observations (with non-policy jobs) were that the scheduler would auto-cancel jobs that had a total execution time exceeding the selection list timout value.  The job would have to complete the backup and verification within this time frame (job wait time+backup time+verify time <= selection list timout).

Example: I configure a job with 14 hour timout that starts at 10pm, the user inserts a tape the following morning at 8am (+10 hours of waiting for a tape), the job would then auto-cancel at noon.

I'm hoping based on the info in the reply above that policy based jobs behave differently.

[pkh]

What am talking about is the time window setting, i.e., how long it will wait before it is re-scheduled.  If you specify the resource availablity limit, then it will also apply when you use the selection list with a policy.  However, you do not need to specify a limit for your resource.  Uncheck that unless the resource needs to be freed for some other purposes.