Forum Discussion
The most common causes of under utilization are;
- Defining more volume pools than you can really fill with data. It is a legacy practice to create ten or hundreds of volume pools in a NetBackup domain. One for Exchange, one for SQL, on for windows, one for unix, etc etc. The problem is that when you do this, each of those pools will take one media and NetBackup will write what ever you tell it on that media. Problem comes when you don't tell NetBackup to write any other backups on that pool (or media) but instead tell it to write in some other pools. Now you have 20 pools if only a few GB written in each pool (or media) that could have been written on one pool (or media)
- Having too many different retention periods and not allowing multiple retentions on a media. If one backup requires 12 months it will take a media, if another backup requires 10 months, it will take another media. I'm referring here to generally same type of backups, i.e all monthly backups, or all weekly backups.
Golden Rule is to reduce complexity in all aspects of your NetBackup environment. That goes for policies, retention periods,volume pools, etc. If you there is no specific requirement to have more than one of ANYTHING do not create more than one.
Examples.
You've got 100 Windows servers in your environment. How many policies do you need?
Answer: 1 (assuming no apps / databases and that you back them up exactly the same, schedule and backup selection)
Only when someone makes a case for having a different schedule or backup selection do you need to create another policy.
The same goes for pools.You can probably get away with Daily, weekly, monthly and quaterly or yearly if there is a requirement for that. But once again, only if there is a requirement.
Hope that makes sense :)
Thank you for the prompt reply,
Its true we have many volume pools, and the paticular one with more un utilized tapes is the one that were use more often for weekly backups. Shows tapes which are not full, active but not in the Robot Control Host, others are last mounted in around 2010 but never wirtten and not in the Robot Control Host.
What is really puzzling, is that tapes which are 800GB getting full at 200GB, and its becoming more regular this year. When no changes have been on pools, retentions(increase in the pool numbers)
- sdo5 years agoModerator
LTO3 native capacity is actually 400GB. You get 800GB only if you are lucky with perfectly compressible backup data. If you backup compressed database dumps to LTO3 then you will never get more than 400GB on a tape. If your backups are slow and cause lots of streaming gaps then you could get as low as 300GB on a tape.
LTO tapes should be good for around 260 full writes/erases by NetBackup, see:
https://docs.oracle.com/cd/E38452_01/en/LTO6_Vol1_E1_D7/LTO6_Vol1_E1_D7.pdf
...so that usually means 260 mounts in NetBackup.
Why 260 mounts, when a tape might be rated for 16,000 full winds and rewinds? Because to fill an LTO tape just once means that the tape will have to fully wind forwards and fully wind backards many many times even just to fill a tape once, due to tape tracks, see:
https://en.wikipedia.org/wiki/Linear_Tape-Open
...yet this old guide suggests 364 full writes for LTO3:
- sdo5 years agoModerator
You say you have tapes from 2010, how many mounts have your problem tapes had ?
Ever lost your catalog, and lost the history of tape mounts ? Maybe your tapes have actually had many more mounts than even NetBackup knows about right now.
- Jim_MD5 years agoLevel 3As others said less pools the better. Another reason is every pool has max number of partially filled tapes attribute. If you have a small number tape drives I tend to do "number of drives + 1". For this to be really effective you need to keep number of retentions limited as well.Be careful if you have reached the max number of partially full tapes backups will queue.
Another thing to check regularly is for tapes that are frozen. Sometimes they are frozen unnecessarily.
- Marianne5 years agoLevel 6
My 2c about some of your statements:
.....Shows tapes which are not full, active but not in the Robot Control Host, others are last mounted in around 2010 but never wirtten and not in the Robot Control Host.
So, this means that these tapes are not in the robot, right?
Do you know where they are so that you can put them in the robot for backups?
NBU will never request tapes for backup if they are not in the robot.
What is really puzzling, is that tapes which are 800GB getting full at 200GB, and its becoming more regular this year.
This is not a NetBackup issue.
Please see this excellent answer from mph999 :
https://vox.veritas.com/t5/NetBackup/The-backup-size-on-the-tapes/m-p/710551#M190323
and https://vox.veritas.com/t5/NetBackup/Data-storing-issue-on-Tapes/m-p/662369#M173226It seems all of your infrastructure is very old - your hardware and software ran out of support many years ago.
Current LTO technology is at LTO8 and NBU version at 8.2.
Maybe your management need to consider upgrades?- mph9995 years agoLevel 6
Thanks Marianne, I tried to post similar last night but it was showing an error.
Here is the technote that explains the same, if you.want something more official.https://www.veritas.com/support/en_US/article.100014480.html
Related Content
- 13 years ago
- 8 years ago