Besides storage device and storage unit NetBackup also has storage unit groups which has got another advantage for efficient backups.
We already discussed about the storage devices such as physical disk and tape storage devices those can be attached to the media servers. We also discussed about the storage units which are logical entities in NetBackup and associated with physical storage devices.
Configuring storage unit groups
When we configure a backup job the NetBackup will request for a storage destination and normally we send a backup to a storage unit. However, we can also send our backups to storage unit groups.
This will let us assign multiple storages belonging to a single group. We also have the option to specify a manner and an order to use for the backup jobs. So if a storage unit is not available then backup job will automatically go to the next available storage unit.
On demand only and any available
Lets discuss about on demand and any available storage unit which are two setting in NetBackup available for earlier versions. We see four different storage units above.
On the left there is basic disk storage unit and on the left there are three tape libraries that are connected to three different media servers.
On the top left side we see a backup Policy1 which has assigned storage “ms1_bdisk_STU” which is our basic disk storage unit. If this backup has to run then it would the same storage unit. Policy2 has any_available it means that NetBackup will try use any storage unit available.
However it is not going to use “ms1_bdisk_STU because it set to “On Demand Only” meaning it would only be used if you point a backup job directly to it. It will however try to use any of the tape storage since they are not set as on demand only.
if we have Policy3 which is pointing directly to one of the 3rd tape library then it would use the 3rd tape library when the backup would run. The function of the “On Demand Only” setting is to prevent that storage unit from being used when a policy goes to “any available storage”.
Note: Even though we have the options to use any available storage it is not a best practice to use it for the backup jobs. Veritas recommends to use storage unit groups to provide multiple destination options for a backup.
Prioritizing storage unit groups
One important thing to understand about the the storage unit group is that there are different selection scheme that cab be used. For example, above we have three tape libraries that is put into one storage group which we can see outlined in red dash lines. On the left we can see the names of the storage units that we have put in the storage unit groups and the order as well.
The default section is prioritized so this is a prioritized storage unit group and what is means that top storage unit would be used first before the second one and second one before the 3rd one.
By looking at the jobs we can find out that job1 and job2 are assigned to 1st storage unit and once the storage unit1 is at its capacity the next jobs would go to 2nd storage unit and next jobs would go to 3rd storage unit.
When all the 3 storage unit are occupied and if additional jobs are added like job 7 and job 8, then these jobs need to wait and would be queued until any of the 3 storage unit is free to accept more jobs. In the above example we have used tape libraries with two tape drives which can handle two jobs at a time.
In a prioritized environment, it would make more sense if all the media severs have different capacity and resources to handle more jobs from one another. So if media server1 has more capacity then we would prioritize the backup jobs to use media server1 before they use other media servers.
What is round robin storage unit groups?
In the next example lets discuss the round robin storage unit groups. In this case also we still have our three tape libraries from three different media servers in one group.
However we notice that when try to assign the jobs, they are getting distributed among the storage units more fairly. The job1 would go to 1st storage unit, the job2 to 2nd storage unit and job3 to 3rd storage unit and like wise the job4, job5, and job6.
So above we have even distribution of jobs to storage units. Once all the storage unit reach their capacity then additional jobs like job7 and job8 are going to queued up and they would go to the next available storage unit whichever one that may be.
This type of storage unit makes more sense in order to evenly distribute the backup jobs and share the load balancing among the media servers.
What is failover storage unit groups?
Here we’ll discuss about another storage unit group type which is failover storage unit group. In this example we have two storage units. From media server1 – basic disk storage unit and media server2 – Tape library.
In this scenario all our backups would go to disk and we want them to go to tapes only if disk storage unit is unavailable. So in a failover storage unit group all the jobs would go to 1st storage unit and once it reaches it capacity then all other jobs would queue up and wait for the same 1st storage unit to accept the queued job but will not go to other storage units.
And the only reason the jobs would go to 2nd storage unit if 1st storage unit is no longer available. This could happen if media server 1 is shut down or file system is unmounted etc.
Also read – How to create NetBackup Policy and schedule