UltraSearch 10g: Moving indexes to new TS, crawler errors, and robot excluding.

  • From: "Jesse, Rich" <Rich.Jesse@xxxxxxxxxxxxxxxxx>
  • To: "ORACLE-L (E-mail)" <oracle-l@xxxxxxxxxxxxx>
  • Date: Wed, 2 Mar 2005 11:42:30 -0600

Hey all,

Anyone here using Oracle's Ultra Search?  The MetaLink forum is =
apparently unmonitored and even Google hasn't helped much.  I've got =
some outstanding issues (at least three) on our 10.1.0.3.0 Ultra Search =
instance running on Solaris9.

First, after successfully crawling our intranet, I see that the =
DR$WK$DOC_PATH_IDX$I table (presumably an Oracle Text object) is growing =
much larger than I had expected.  While there's no immediate need, =
should I want to move this to another tablespace, how?  I seem to =
remember that Oracle does not support the movement of some Oracle Text =
objects, but I'll be darned if I can find info on this.

Second, I'm getting some crawler errors in the logs.  It started off =
with an ORA-00001 on WK$CACHE_PK and has seemingly turned into =
"ORA-01003, no statement parsed" since.  MetaLink says the ORA-1 is =
fixed and there's no articles on the latter.  Methinks not.

Third, I'm using a robots.txt file to exclude the web paths to be =
crawled, but since our intranet uses frames (despite my insistance that =
frames are evil, like the fruits of the devil, evil) the navigation =
frames are being included in the crawl.  I've tested using the ROBOTS =
META tag to disregard indexing the navbar, but follow it's links, but it =
looks like Ultra Search ignores those.  Anyone successfully use a =
different method?

If this is like my previous Ultra Search posts, I won't hold my breath =
for an answer...

TIA anyway,
Rich

Rich Jesse                        System/Database Administrator
rich.jesse@xxxxxxxxxxxxxxxxx      QuadTech, Sussex, WI USA
--
//www.freelists.org/webpage/oracle-l

Other related posts: