IBM Support

MustGather: Troubleshooting problems with WebSphere DataPower SOA Appliance configuration objects

Technote (troubleshooting)


What is the best approach for debugging problems with configuration objects that appear to have been lost, not saved, or are missing after some event?


At any given time, two versions of your DataPower configuration exist.:

  • The running or "in-memory" configuration consists of objects created via import or by applying changes, but which has not yet been saved.
  • The startup or "persistent" configuration consists of those objects which have been persisted via "Save Config" in the WebGUI or "write mem" from the command line interface to the configuration file within config:///.

    NOTE: Only objects in the configuration file are loaded after an appliance reboot or reload.

The appliance includes a Configuration Comparison utility to determine what has changed between configurations. This can be found in the Objects>Compare Configuration panel.

When there are unexpected or unexplained problems with configuration objects, there are some logged events of specific types and categories which may be helpful in problem determination. Typically, the most useful messages pertaining to configuration issues will be logged as "cli", "conf", or "sys" category events.

Logged Event Categories Sample Message Comments
cli 1,20110428T162551Z,default,cli,



configure terminal"

User login- logged at debug logging level
conf 20110422T142712Z [conf][failure] (SYSTEM:Prod:*:*): certificate 'MyNewCertificate' -

File is unreadable

Indicates a problem with an unreadable certificate file that can affect the status of configuration objects dependent upon it:
sys 20110413T214123Z [sys][success] (SYSTEM:default:*:*):

Booting build 190083 on 2010/09/20 13:56:21 count 1. Uptime 65

Confirm whether a restart may have affected the running or non-persistent configuration:

These events are very important because, as noted in the examples above, confirming that a user logged in, configured or changed a configuration, and then the device was reloaded or rebooted may be key to understanding the observed behavior.

Along with the events logged, the autoconfig.cfg file is another key piece of information which can be found in the config: directory. This is the persistent configuration that is updated when the configuration is saved in the WebGUI or a write memory is performed.

Diagnosing the problem

If you have examined the logs, but are still unable to determine the cause of the observed behavior, collecting the following artifacts before opening a PMR with IBM Support can help expedite problem resolution. Read through the following steps to plan for capturing data at various stages during a recreate attempt:

1) Configure a reliable log target in the default domain with the event subscriptions as seen in the following screen. Ensure it is active and logging to the RAID Array hard drives or another location if there is no other choice. It is recommended to avoid logging to the logstore: directory to avoid risk of exhausting file system space.

    2) Collect the following basic information. See MustGather - Collecting data: WebSphere DataPower SOA Appliances problem determination

    3) Collect the following files. If possible, collect this data before & after saving the running configuration to the persistent configuration:

    • If the problem occurs in the default domain, collect file config:\\\autoconfig.cfg
    • If problem occurs in non-default domain, collect file config:\\<domainname>.cfg

    Also collect files: audit:\\\audit-log, audit:\\\audit-log.1, and logtemp:\\\cli-log

    Then contact IBM DataPower support with the information collected.

    Resolving the problem

    Document information

    More support for: WebSphere DataPower Integration Appliance XI50

    Software version: 3.8.2, 4.0.1, 4.0.2

    Operating system(s): Firmware

    Software edition: All Editions

    Reference #: 1497973

    Modified date: 21 July 2011