McIDAS-XCD Administrator's Guide
Version 2010.1

[Search Manual] [Table of Contents] [Go to Previous] [Go to Next]


Chapter 4

Troubleshooting

This chapter lists problems that may occur with McIDAS-XCD. Under each symptom or error message, possible solutions are given. The problems discussed here are listed below.


-XCD is not receiving real-time data - users report no real-time data, or the ingestor status display is red


-XCD is not creating McIDAS Grid files

The DMGRID GRIB decoder can't file grids if it can't find RTMODELS.CFG, which contains information about real-time grid file locations.

The file RTMODELS.CFG should reside in ~oper/mcidas/data when McIDAS-XCD is installed correctly. Either the decoder can't reach the file or it is missing. If it's missing, recreate the file or copy a new version of the default file from ~oper/mcidas/xcdversion/data/RTMODELS.CFG (where version is the version number, e.g., 2008).


-XCD is not writing GRIB files

From a McIDAS-X session running under the oper account, run the REDIRECT LIST command to check for any file redirections.

Type: REDIRECT LIST

Look for the *.gr* and *.bu* directory entries and verify that these directories exist on your workstation.  If these directories do not exist, create them. Once this is done, -XCD will automatically start writing GRIB1 and GRIB2 files to this directory, which in turn will allow for data to be written to the database. Then users should be able to access the GRIB data.


-XCD is not writing to the real-time GRIB MySQL database

From a Unix shell prompt under the oper account, run the gribadmin command to check the most recent record filed in the mcrtgrib database.

Type: gribadmin latest


-XCD data is garbled or missing

When more than one ingestor is trying to read the same circuit, text data may be missing or text output garbled. If you are decoding grids, grids may be missing.

Only one ingetext.k process should be running for each text circuit, and only one ingebin.k process should be running for each binary circuit (e.g., one ingebin.k process running for GRIB data while another is running for NEXRAD data). Check the number of ingetext.k and ingebin.k processes and -XCD data monitors running. At a Unix shell prompt, as user oper,

Type: ps -ef | grep DM

Type: ps -ef | grep inge

If there are multiple processes running for one of the circuits or one of the data monitors, stop and restart -XCD and the LDM. Some data may be lost while the processes are stopped.

  1. Stop -XCD and kill any remaining data monitors (DM*), if necessary. At a Unix shell prompt, as user oper,
  2. Type: xcdadmin stop (you will need to wait for the data monitors to cleanly exit)

    Type: ps -ef | grep DM

    Type: kill -9 process_id (if necessary)

  3. Stop the LDM and kill any remaining processes, if necessary. At a Unix shell prompt, as user oper,
  4. Type: ldmadmin stop (stops the INGEBIN and INGETEXT processes)

    Type: ps -ef | grep inge

    Type: kill -9 process_id (if necessary)

  5. Restart the LDM and -XCD by running the two commands below, in order.
  6. Type: ldmadmin start

    Type: xcdadmin start

If this process doesn't work, check for an obstruction in the receiving antenna.


xcdadmin script is not working


directory ***NEXRAD_PATH*** exists on the workstation

The ***NEXRAD_PATH*** directory is created if the NEXRAD data monitor is enabled and the DIRFILE= value is not set in the file ~oper/mcidas/data/NEXRID.CFG. See Chapter 2's Configuring the NEXRAD Decoder section for instructions to set the DIRFILE= value and other NEXRAD options.


[Search Manual] [Table of Contents] [Go to Previous] [Go to Next]