We have moved to www.dataGenX.net, Keep Learning with us.

Tuesday, August 13, 2013

Error 39202 when logging into IBM DataStage


DataStage Client programs (Designer, Manager, Director) experience a connection failure, and the message displayed to the user is:

Failed to connect to host: xxxxx, project: UV
(Internal Error (39202))

Solution :

This error indicates that a problem on the DataStage engine has caused the server dsapi_slave process to fail, and the connection is lost. There are a number of different possible root causes for this message to be displayed, but the most common cause is file permission problems.


There are several possible causes for this error. Go to the $DSHOME directory and run:
$. ./dsenv
$bin/uvsh

If this fails it points to a problem with the engine. If it works correctly and you get the TCL prompt you may have a hung socket. Perform the following steps to cleanup DataStage.


  1. Log into server as dsadm or root.

  2. Make sure the user has permissions to /tmp and that the partition is not full.
  3. From $DSHOME type:
    $. ./dsenv
    $bin/smat -t


    Check that the user has permissions to the location for UVTEMP and UVSPOOL.
  4. If you made any changes to ODBC drivers check the permissions under $DSHOME/../branded_odbc.
  5. Stop DataStage and check for orphaned processes:

    $ps -ef|grep phantom
    $ps -ef|grep dsapi
    $ps -ef|grep dscs


    Use kill -9 for any existing processes

  6. Check for existing shared memory segment:

    $ipcs -mop|grep ade (Linux users - use ipcs -m)

    As root remove any existing segments:

    $ipcrm -m PID

    Example:
    # ipcs -mop |grep ade
    m 13500432 0xadec7011 --rw-rw-rw- root dstage 0 88240 88100

    #ipcrm -m 13500432


  7. Check for hung ports:

    $netstat -a|grep dsrpc

    If you get any results in a WAIT state then you need to wait until this it disappears or reboot the server. Depending on the exact state it may clear itself or may never go away. For example a TIME_WAIT is controlled by the ndd parameter tcp_time_wait_interval, which is generally 60s.

  8. Start DataStage and Re-connect with DataStage clients.