Archives of the TeradataForum

Message Posted: Mon, 21 Apr 2003 @ 11:02:38 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Here is my response to one readers concerns regarding this method:

The script creates two processes. They can be anything. In this case the are gzip and mload. Generaly process A should write to the pipe and process B should read from it. This script is most handy when you have multiple zipped data files to load.

You create a pipe with "mknod name p". You need to read about named-pipes if you are unfamilar with how they work. Try "man mknod" or look on groups.google.com. There are tons of material available there.

The only problem I've seen are: 1) if process A bombs out before the monitor.pl script has a chance to start it's monitor loop then it could just wait forever. In my experience this happens mostly while you are developing the code. Once it is working you should experience this infrequently.

We use this script here on a daily basis with well over 200 mloads, fastexports, sqlldr, etc. So I can attest for the reliabilty of the script. You will need to write some wrapper code to check you return codes once monitor.pl exits.

	<Prev	Next>		<<First	<Prev	Next>	Last>>