cancel
Showing results for 
Search instead for 
Did you mean: 

BE 20.5 maximize tape usage

KPS_BV
Level 3

We have:

- one single media server

- one library with 8 slots and 1 LTO6 drive

- We run a daily job to tape for incremental and a weekly job for full backup.

- Each full backup is less than 500 GB. So, 4 full backups fit onto single weekly tape.

- We use 4 weekly tapes and we can ensure that only one single weekly tape is in the library

 

Target: We want to use a different tape in each 4-week cycle to maximize tape usage.

Week 01 to 04: backup is written to corresponding tape 1 – 4 (first backup one each tape)

Week 5 to 8: backup is written to tape 1 – 4 (second backup appended on each tape)

Week 9 to 12: backup is written to tape 1 – 4 (third backup appended on each tape)

Week 13 to 16: backup is written to tape 1 – 4 (fourth backup append each tape)

Append period ends after fourth backup to tape and cycle starts again:

Week 1 to 4: backup is written to tape 1 – 4 (1st backup one each tape) because overwrite protection period has ended.

 

To achieve this we defined a media pool “weekly” and assigned to following parms:

Append period: 13 weeks

Protection period: 3 weeks

Would that fulfill our expectation as described above?

Thanx for any feedback 

2 REPLIES 2

Gerhard-S
Level 3

Hi,

at first: the APPEND period is a kind of usage-window, meaning as long as append is possible the RETENTION (overwrite) timeframe does not start.

the append time should be as short as possible to have control over the media-usage.

secondly: to put more than a week on a single media is dangerous, because think about the consequences if in week four the write fails because of a "mechanical" issue, you'll loose multiple weeks of data. Each writing access raises the risk of losing (yes, i'm paranoid, but why push your luck?).

third: you can partition the library to divide the available slots for different tasks, meaning you could mark the slots for the operators clearly and use only specific slots for specific jobs see https://www.veritas.com/content/support/en_US/doc/59226269-137906722-0/v53896777-137906722

regards

Gerhard

Thanx. I agree to your general risk analysis. But I think our risk weights less because we plan to write only every 4th backup to the same tape so in the case of tape damage there will be 3 consecutive backups on different tapes available. We change the tape after every full backup.