IC76189: AGENT REGISTRATION FAILS ON CLONED IMAGES/SERVERS HOSTING DATA AGENTS/SRAS DUE TO DUPLICATE GUIDS

4.2.1 Interim Fix 1 (November 2011) for Tivoli Storage Productivity Center

APAR status

• Closed as program error.

Error description

• The GUID is used as unique Identification for an agent, thus
  when a TPC Server gets data from an Data Agent/Storage Resource
  Agent we use unique GUID to identify the agent and save its
  respective data.
  
   This means that if there are multiple agents with same GUID
  they just keep overwriting the data from each other.
  
  This problem gets manifested by fact that deployment of machines
  and/or VMs using image replication is very common practice.
  Hence we can expect quite a few customer to run in to issue. We
  publish a way force regeneration of GUID for such situations,
  which need to be followed before the master imaged is built.
  i.e.
  
  TPC Hints and Tips Guide v.4.1.x
  
  http://www-01.ibm.com/support/docview.wss?uid=swg27008254&aid=1
  
  Refer to: Section 5.2 'Creating a master image to clone TPC
  Agent machines'
  
  Which explains how to correctly create a master image used for
  cloning, thus, eliminating the duplicate TIVGUID issue.
  
  However, this procedure is often not used because the personnel
  building the OS images are not supporting TPC.
  
  To protect TPC from overwriting agent information we need a
  better protection in place. We need to modify our agent
  registration process as follows:
  
  1. At agent (SRA/Data agent) registration time by default TPC
  server must not allow an agent to register if the an(another or
  same we can not distinguish) agent already registered using same
  GUID.
  
   ( Note that this check should be done at new registration time
  only. Probe and scan etc should continue working with current
  logic of overwriting info using the GUID as unique identifier. )
  This would it turn fail the SRA installation on the machine
  which try to register with that GUID.
  
  2. As  we still need to cover for backup/restore and restoration
  of image to same machine, we need to govern the behavior 1 with
  a config option. The default behavior should be as above, but
  when the config option is set otherwise, we need to allow new
  registration to overwrite existing agent (the current behavior).
  
  Problem as described by customer:
  Duplicate GUIDs cause DA/SRA previously registered to the TPC
  Server with the same GUID to lose connectivity to TPC Server.
  
  Initial customer impact (low/med/high):
  High
  

Local fix

• None at This Time.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: Users cloning server machines with GUID      *
  *                 preinstalled.                                *
  ****************************************************************
  * PROBLEM DESCRIPTION: Since TPC indetifies the cloned         *
  *                      machines with Storage Resource Agents   *
  *                      uniquely based on GUID, so only one     *
  *                      with last probe remains in TPC          *
  *                      overwriting other, with no error given. *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  -
  

Problem conclusion

• Change the registration to check if a server is already
  registed with same GUID and the host name is not same, Now TPC
  will not register the new one and give error.
  
  The fix for this APAR is targeted for the following maintenance
  packages:
  
  | fix pack | 4.1.1-TIV-TPC-FP0007 - target to be determined
  | fix pack | 4.2.1-TIV-TPC-FP0004 - target August 2011
  
  http://www-01.ibm.com/support/docview.wss?&uid=swg21320822
  
  The target dates for future fix packs do not represent a formal
  commitment by IBM. The dates are subject to change without
  notice.
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  TPC FOR DATA

• Fixed component ID

  5608TC300

Applicable component levels

• R41A PSY

     UP

• R41H PSY

     UP

• R41L PSY

     UP

• R41S PSY

     UP

• R41W PSY

     UP

• R42A PSY

     UP

• R42H PSY

     UP

• R42L PSY

     UP

• R42S PSY

     UP

• R42W PSY

     UP