OPTIMIZE. A selection of projects which have significantly improved customer environments.

Storage Graph

Optimizing storage layout for backup and replication

Occasionally designs need to be re-factored, this usually occurs when the requirements of the original design change.

Designing a greenfield solution allowed us to generate a migration strategy making use of qtree SnapMirror (QSM), and object rename/moves. A name accurate environment was created which was used to validate the scripted migration process and its step-wise back out processes. Once validated the entire process was handed to the customer’s operations team to execute.

The ability to map out and validate the migration process without impacting the customer’s environment ensures that migrations are successful the first time.

NetBackup Tape usage visualization

One feature of HP OmniBak software is the ability to visualize tape drive usage. While the data to report on tape drive utilization exists within NetBackup the ability to easily and quickly view this data does not yet exist.

Using the bptm logs we are able to generate a time line of tape mount/unmount requests. Layered over this we are able to map in jobs being multiplexed onto the same tape.

Visual representation of drive use and NetBackup multiplex generated graphs allows us to:

  • Quickly identify multiple incarnations of the same retention policy which prevents multiplexing from working correctly
  • Provide high frequency tape drive mount/unmounts due to Oracle archive log backup and a short media unmount delay
  • Shared Storage Option (SSO) issues
  • Provide easy to digest information to capacity planners

Being able to quickly identify and visually represent issues within customer environments improves return on capital by increasing effective capacity, and improve overall backup efficiency and backup success rates.

NetBackup tape multiplex by time graph
Graph of a Sun system's PCI bus utilization

Performance Analysis

In terms of business impacts system performance ranks second only to system availability. While it is possible to design a degree of availability into a solution it is harder to design one with respect to performance. This becomes especially difficult when the performance requirements of the solution are unknown or are rapidly evolving. When this happens we have the ability, tools and process to determine the root cause and synthesise a solution.

Identified issues include:

  • TCP windows causing performance related service outages within NetBackup and NetApp
  • Limitations with IBMs ATape driver when used with NetBackup
  • Suboptimal SQL queries and database layouts
  • Configuration issues between storage/volume manager/file system layers generating 4 times the IO load of a correctly designed system

In all cases the issue is confirmed and triaged, the root cause identified and required actions recommended to minimize or reduce the issue being experienced.

Contact information

Copyright 2013 Unicity P/L