Sign In
Problem
When configuring a Vault Profile to run (only) duplication jobs of backup images written to an OST Disk Storage device shared by multiple media servers; the Vault Duplication > Destination > Write Drives was set to 6; however, when running test vault duplication jobs, not all of the 'write drives' are being used for the vault duplication jobs (only 2 out of 6 for initial tests).
For the customer reporting this problem; their end-goal was to use all 6 configured vault duplication 'write drives' concurrently to complete the duplication of large amounts of data from the OST Disk Storage device to tape as quickly as possible.
Error Message
N/A
Cause
When running vault duplication jobs; the number of tapes/tape drives used for the duplications depend on the following criteria:
1. The value (i.e. number of write drives) specified from the Duplication (tab) > Destination > Write drives:
2. The number of DIFFERENT retention values of the original (primary) copies of the images 'Chosen' to be duplicated.
3. The policy configuration (in this case, the use of a load-balanced Storage Unit Group as the 'Destination > Policy Storage' for all primary backups)
When a Vault Duplication job is started, the following happens:
* NetBackup checks the backup images that meet the Vault Profile's > Choose Backups specified criteria to select the images to be duplicated
* A list of all backup images that do not already have a 'copy 2' and that meet the vault profile's 'Choose Backup' criteria is created
* NetBackup then separates the images in the master list into separate 'batch' files that are run as individual duplication jobs to a single tape/tape device
Note: The number of tapes/tape drives used depends on the number of batch files created; and the number of batch files created is determined by the number of DIFFERENT retention levels are found in the backup images to be duplicated; and, when a shared OST intelligent disk device is the 'Source' for the vault duplication jobs, separate 'batch' jobs are also created for each media server that write to it.
Solution
Use the formula (# of retention levels X # of media servers = # of batch jobs/tape drives used) to determine how many tape drives would be used per profile; then configure multiple vaults (each with its own single profile) to run concurrently which will force the use of all available vault duplication 'write drives' for optimal performance of vault duplication jobs.
Applies To
Note: All names listed below are generic names listed to clarify examples in this document.
NetBackup 7.x Environment:
-------------------------------------------
* Master Server: Master-bn1 - Solaris 10 - running NetBackup 7.x
* Media Server: Media01-bn1 - Solaris 10 - running NetBackup 7.x
* Media Server: Media02-bn1 - Solaris 10 - running NetBackup 7.x <- Alternate Read Server for all Vault Duplication jobs <- Zoned for 8 tape drives
* Removable Storage: TLD(4) - Robot: IBM 3584L22 (8900) | Tape Drives: 14 IBM 3592E05 (EC7) | Drive Type: hcart2
* Disk Storage: OST_DD_001 | Disk Type: Open Storage (DataDomain) | Media Servers: Media01-bn1 & Media01-bn1 | Disk Pool: OST_DD_001-dp
Current Vault Configuration: <- For Vault Duplication jobs ONLY!
----------------------------------------------------------------
Choose Backups:
- Choose backups between: 13 days and 0, 0, 0
* Attributes:
- Clients: <All clients>
- Backup Types: <Include all backup types>
- Media servers:<Include all media servers>
- Backup policies: <Selected a Single Policy> <- This is the only distinguishing criteria that determines which backup images are duplicated.
- Schedules: <Full_Schedules_Only>
- Retention level : <Include all retention levels>
* Locations:
- Disk Pools: OST_DD_001-dp
Duplication:
- Source backup reside on: Disk Only
- Number of read drives: 6
- Alternate read server: Media02-bn1
- Storage Unit : Media02-bn1-hcart2-robot-tld-4
- Volume Pool : Off_Site
- Retention: 7 years
* Catalog Backup: Skipped
* Eject: Skipped
* Reports: N/A
Note: At the site described above; ALL policies have been configured to use a 'Load-Balanced' Disk Storage Unit Group as the 'Policy storage' selection for all backups; and that Disk Storage Unit Group included the Disk Storage Units configured for the shared OST storage device for both media servers at this site. As a result, the single policies selected as the distinguishing criteria on the vault profile's 'Choose Backups" tab, are run from both media servers at different times.
Terms of use for this information are found in Legal Notices.
