SDI Operator's Manual
Revised July 2000

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

Appendix B

Frequently Asked Questions

This appendix is a list of questions that the SSEC Data Center commonly receives from SDI users.

Q. I just started my new NOAAPORT SDI. Why does McIDAS-XCD quit processing data and display the error : spool files wrapped and data lost on the console?

A. This error means that nothing is reading the FIFOs. Make the following checks.

Q. How do I remove a "bad" image (e.g., an image with a future date and time that was generated due to a problem at the ground station) from my GVAR SDI or POES SDI?

A. First, remove the index files for the bad data from the data directory (usually /data). See GVAR Image Index Files in Chapter 3, or POES Index Files in Chapter 4 for more information about index files.

Then remove all references to the bad image in the descripter files. Be careful not to edit the descriptor files while image starts are writing to the files. Edit the files between images, or stop the ingestor, edit the files, then restart the ingestor. See GVAR Descriptor Files in Chapter 3, or POES Descriptor Files in Chapter 4 for more information about descriptor files.

Q. What is a FIFO?

A. FIFO stands for First In First Out. It is a pipe that looks like a file.

Q. Why does my new NOAAPORT SDI / McIDAS-XCD system quit seeing new data after running for only a few hours?

A. Chances are that /tmp is full on your NOAAPORT machine because the ingestor is writing more 1MB spool files than /tmp has room for. If this is the problem you have two options:

Q. I have just edited the /etc/init.d/ingcntl file, but when my SDI rebooted, the changes did not take affect. Why?

A. The ingestors use /etc/rc3.d/S99inge when they boot to start the ingestor. This file should be hard linked to /etc/init.d/ingcntl, so that any changes made in one will also be in the other.

Q. What is the difference between files /etc/init.d/ingcntl and /etc/rc3.d/S99inge?

A. S99inge is a script in the system run-level directory /etc/rc3.d. The file /etc/init.d/ingcntl is the ingestor startup script, used mainly for manually stopping and restarting the ingestor.

In order for the ingestor to restart automatically upon system booting, an ingestor auto-startup script must be included in the system run-level directory /etc/rc3.d, where run-level 3 is the normal multi-user level. When SDIs are shipped, /etc/rc3.d and /etc/init.d/ingcntl are hard linked (so all changes made in one are also made in the other) so the ingestor will be brought up when the system reboots.

Q. Can both the ingestor and server be installed on a NOAAPORT SDI that is not running -XCD?

A. Yes they can. But the server software will not be used unless the signal type is switched to run as a different type of ingestor (e.g., GVAR).

The only time the servers should not be installed on a SDI is when McIDAS-XCD is also installed. The reason is that if the ingestor servers are installed they will override the McIDAS-X and -XCD servers (and the -X and -XCD servers are the ones that should be used for serving NOAAPORT data).

Q. I've configured my NOAAPORT ingestor to ingest GOES/GINI data in area format, but no areas are being created. The ingestor console is showing the error makgini: cannot make positive UC: could not create nnnn byte shared memory segment. What's wrong?

A. This usually indicates that your shared memory is not set up correctly. Verify that the line listed below is in the file /etc/system. If it isn't, add it, then reboot the system.

 set shmsys:shminfo_shmmax = 0x20000000

Q. Why does the message unable to write to utmpx appear when I stop my SDI ingestor?

A. Nothing is wrong. This message is normal when the ingestor shuts down.

Q. Can I feed two McIDAS-XCD servers off of one NOAAPORT SDI?

A. Yes. See Configuring Two McIDAS-XCD Workstations to Feed From a Remote SDI in Chapter 6, NOAAPORT Ingestor to find out how.

Q. I have two McIDAS-XCD servers set up to feed off of one NOAAPORT SDI, but only one server is getting data. What's wrong?

A. Each -XCD server must use its own pair of ports (1501 & 1502 or 1503 & 1504). For example, only one server will see the data if you are using port 1501 for one machine and 1502 for the other, or if you are using port 1501 for one machine and 1503 for the other but you don't set ports 1502 and 1504 to /dev/null using the Unix command cat. For more information, see Feeding off Multiple NCEP/NWSTG Data Streams in Chapter 6, NOAAPORT Ingestor.

Q. Can the GVAR SDI be configured to hold, for example, more CONUS images than SH images?

A. No. It can't because the GVAR SDI retains data in the order it is received. We suggest writing images you want to keep to another local machine.

Q. We just installed our SDI, but upon startup, we get the message Open error on /dev/jmb no such device or address, and no data is ingested. Why?

A. This message indicates that the SDI board was not found by the ingestor. Reseat the board.

Q. What time zone should my SDI ingestor be set to?

A. The time zone setting does not matter to the SDI.

Q. Do you need a separate SDI machine for each NOAAPORT signal type (e.g., NWSTG channel and GOES channel)?

A. No. But you do need separate SDI cards. SSEC sells 2-card SDIs for ingesting the NCEP/NWSTG channel and one of the GOES channels.

Q. I do not want to file AREAs on my NOAAPORT-GOES SDI, but areas are filing. How do I shut off the AREA production?

A. Be sure that no files named /opt/nport/area.TIGnnn exist. If they do, delete or rename them all.

Q. How do I increase the number of GINI files being saved?

A. To lengthen the loop of GINI files being saved, add the file type and number of files to save to the file /opt/nport/retain.nport.

Q. How do I monitor my NOAAPORT text datastream?

A. See the examples in the Chapter 6, NOAAPORT Ingestor sections titled Feeding off the NCEP/NWSTG Data Stream Locally or Feeding off the NCEP/NWSTG Data Stream Remotely. Note that you should use the unused FIFOs or unused ports. For example, if FIFOs 1 and 2 and the corresponding ports 1501 and 1502 are being used by McIDAS-XCD, then FIFOs 4 and 5 and ports 1503 and 1504 are available for monitoring.


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