> On 06Oct2009 08:14, Yong Huang <yong321@xxxxxxxxx> wrote:
> | I have RHEL 4 or 5 with standard installation. I need a compressed file
> | system temporarily to store an Oracle data pump export file. What's the
> | easiest solution? Oracle data pump does not support a sequential device
> | as its output so I can't pipe the output to gzip or compress.
>
> Does it actually do random I/O to the dump file, or does it simply
> require you to specify a filename with no _convenient_ pipe output
> syntax?
>
> If it is only an invocation issue, thus:
>
> oracle-data-pump-command filename
>
> you may be able to check and use a named pipe:
>
> mknod p foo
> gzip <foo >foo.gz &
> oracle-data-pump-command foo
Oracle's official document says data pump export does not support pipes as output (unlike the conventional export for which people usually create a named pipe as you showed). It probably does random I/O to the dump file.
I researched some options for natively compressed file systems for Linux. I just want to know which one people think is the easiest. Not a big deal. Squashfs.sourceforge.net looks promising. Thanks.
Yong Huang
--
redhat-list mailing list
unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/redhat-list
