IBM Support

IV65154: IBM Endpoint Manager 9.0 client consumes all of the CPU and become unresponsive.

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as program error.

Error description

  • From the logs we can see that there are a couple of reboots
    happening in a short amount of time. After the first reboot the
    Agent started normally and began to re-process the files
    pre-cached by the multiple action group which the agent had
    received a couple of days earlier. This is normal and it is
    done to verify and reorganize the cache. During this
    processing, the agent received a new stop command due to the
    second reboot request. The stop arrived at a time when the
    agent was still dealing with the cache files and had its
    content in an inconsistent state. Having been interrupted in
    the middle of the process and being the cache in an
    inconsistent state, the agent, before exiting, started the
    recovery procedure that ended up in the infinite loop. Because
    the reboot had to complete, the agent process, hung in the
    loop, was killed. At restart, the agent tried again to clean up
    its cache ending again in the code that causes the infinite
    loop.
    

Local fix

  • In order to recover the client, it is necessary to execute the
    procedure below:
    1. stop the IBM Endpoint Manager client process
    2. delete the <Client Install
    Path>/__BESData/__Global/__Download folder from the IBM Endpoint
    Manager Client cache
    3. restart the IBM Endpoint Manager client process
    
    Or upgrade to IBM Endpoint Manager clients to 9.1 version.
    

Problem summary

  • Under some specific conditions the IBM Endpoint Manager 9.0
    client may enter an error loop. In such a situation, the client
    starts using all of the CPU's resourcess and becomes
    unresponsive.
    While performing the download of a file or checking the status
    of a pre-cached file, the IBM Endpoint Manager client creates
    some metadata files in its cache.
    If an external event triggers the IBM Endpoint Manager client
    stop procedure in the exact moment the client is in the process
    of writing the metadata files, the client may enter into a loop
    while attempting to clean up the partial data. At the next
    start-up the client will enter again into the same loop while it
    tries to recover from the same error condition.
    

Problem conclusion

  • Fixed in IBM Endpoint Manager 9.1 Clients
    

Temporary fix

Comments

APAR Information

  • APAR number

    IV65154

  • Reported component name

    TIVOLI EP MGR R

  • Reported component ID

    5725C4300

  • Reported release

    91W

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt

  • Submitted date

    2014-09-23

  • Closed date

    2014-09-24

  • Last modified date

    2014-09-24

  • 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

    TIV EP MGR AGEN

  • Fixed component ID

    5725C43AG

Applicable component levels

  • R91W PSY

       UP

[{"Business Unit":{"code":null,"label":null},"Product":{"code":"SSBQVS","label":"Tivoli Endpoint Manager"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"91W","Edition":"","Line of Business":{"code":"","label":""}}]

Document Information

Modified date:
24 September 2014