Asset ID:	1-75-1022218.1
Update Date:	2011-07-18
Keywords:

---

Solution Type  Troubleshooting Sure

Solution  1022218.1 :   VTL - What questions to ask and what data to collect for performance related issues  


PreviouslyPublishedAs
277610

Applies to:

Sun StorageTek VTL Prime System - Version: 1.0 - Build 1813 to 1.1 - Build 2076 - Release: 1.0 to 1.0
Sun StorageTek VTL Storage Appliance - Version: 4.0 - Build 1221 to 4.0 - Build 1221   [Release: 4.0 to 4.0]
Sun StorageTek VTL Value System - Version: 1.0 - Build 1323 to 1.0 - Build 1323   [Release: 1.0 to 1.0]
Sun StorageTek VTL Plus Storage Appliance - Version: 1.0 - Build 1323 to 2.0 - Build 1656   [Release: 1.0 to 2.0]
All Platforms
.
***Checked for relevance on 18-07-2011*** (dd-mm-yyyy)

Purpose

This article outlines what questions to ask and what data to collect to start troubleshooting VTL performance issues.

Last Review Date

July 19, 2011

Instructions for the Reader

Troubleshooting Details

Below lists questions and data to collect for possible VTL or Backup App throughtput performance issues:

1) Is VTL patch level current? 
* Recommend to get current, if not already, before diagnosis.

2) Is Backup client OS, Backup application and client HBAs at current code levels?
* Recommend to get current, if not already, before diagnosis.

3) Has anything changed recently in the environment? 
* Patches, FW, workload, backup policies...

4) Is this only occurring on certain types of backups or clients?
* Get example client host names, date & time of a backup and tapes numbers for backups that did not perform to expectations 

5) Any other processes running at the same time?
* Are exports/migrations running concurrently (recommended to have separate windows for backups and exports)

6) Can the situation be reproduced or is this a one time event?

7) If it can be reproduced, throughput numbers and logs will need to be gathered during backups to determine if bottleneck exists (on backend disk, on VTL, on SAN/switch, and backup client hosts). Refer to Knowledge doc id 1015838.1.

8) To determine the severity of a possible performance issue, ask for previous throughput numbers before the slowdown.  Or what is the expected performance and has this ever been achieved?

9) Tips and other areas for performance bottlenecks:
    * Review "VTL Plus 2.0 Update 1 White Paper" on http://download-adc.oracle.com/archive/cd_ns/E21126_01/
    *Verify VTL is spreading I/O across both controllers (e.g., even LUNs on controller A and odd LUNs on controller B)
    * Verify backup app is scheduling backups to spread workload across LUNs (this is a manual process to see which virtual tapes are being used for each backup and what LUNs those tapes correspond to).
    *Verify backup app buffer sizes are set optimally (contact with backup app vendor support, i.e., NBU = 256KB).
    * Verify HBAs, switches, ports, network are capable of handling the desired performance levels.
   

NOTE:  Performance issues are best worked during the normal hours of the customer and local support, as more information is going to needed. Assign case/task to the GEO that the customer is in.

Also, performance issues may need to be handled by professional services, if after initial review is done and no errors or problems are found. Remote support primarily handles break/fix and performance tuning falls under Professional Services area.  Contact Oracle local support in this case.


Attachments

This solution has no attachment