If you want historical tracking of this, cron this to run at a specific
interval. Redirect the output to a file somewhere. Externally map the
directory as an oracle directory and make the file an external table.
Then create yourself history table and as the file updates kick off a
proc in the db to load your history from the updated file. I'm actually
working through a process like this right now to provide some trending
data on disk utilization for our Exec VP's.
Mark W. Farnham wrote:
>
> What I would do is ls –lR every day to a date named file and then diff
> any two days and pick off the rows and fields I wanted and net up the
> ins versus the outs with awk or perl.
>
> Then if you want to drill in on more details about the changes, you’ve
> got them. If your arbitrary time granule is smaller than a day, then
> you’d need more frequent snaps, and of course this process dies by
> Xeno at some level of willingness to spend space for the data.
>
>
> <http://www.orawin.info>
>
--
http://www.freelists.org/webpage/oracle-l
Received on Wed Mar 21 2007 - 09:07:09 CDT
