Named Pipes

named pipes allow one to do such things as redirect data exports straight to a tape device, or to another load command. You can dump from one productline and immediately load into a second productline

Here is a Lawson procedure for dumping with named pipes:

1. Create a fifo file

   mknod named.pipe p

2. Start the data load process. It will wait for input until the end of the process.

   impexp -f productline named.pipe

3. Open a second Lawson session.

4. Start the unload process. This will write to the fifo file which in turn, routes the data to the load process.

   expsysdb -s productline named.pipe

Dump and compress

Here is a procedure that dumps and compresses data from an Oracle export:

1. mknod /tmp/exp_pipe p

2. compress < /tmp/exp_pipe > productline.dmp.Z &

3. exp file=/tmp/exp_pipe userid=username/password full=y

To import the data back in, you would need to uncompress:

1. mknod /tmp/imp_pipe p

2. uncompress < /tmp/imp_pipe > productline.dmp.Z & imp file=/tmp/imp_pipe userid=username/password

Dump and split

Here is a procedure that dumps data and splits it into files that are a maximum of 2GB in size:

1. mknod expsysdb_pipe p

2. split -b2047m expsysdb_pipe expsysdb.dmp_ &

3. expsysdb prodline expsysdb_pipe

This will produce a series of files called expsysdb.dmp_XX, where XX starts out at "aa", then progresses through "ab", "ac", etc,...

To pull the data back in run:

1. cat expsysdb.dmp_* > expsysdb_pipe &

2. impexp prodline expsysdb_pipe

Home