database integrity while indexing

User avatar
Thunderstone
Site Admin
Posts: 2504
Joined: Wed Jun 07, 2000 6:20 pm

database integrity while indexing

Post by Thunderstone »




some users complain that our webinator search database is not accurate
while we are indexing out server. im aware that the -index option (which
we use) causes gw to organize a particular database after indexing ,
but is there a way to make sure that webinator always works correctly when
tables are being built ? we are using the commercial version of webinator
for solaris 2.5.1 .

thanks!,

//Michael.


~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^~`^
Michael Jastremski )( mjastrem@thunder.temple.edu
Http://newsroom.temple.edu )( http://thunder.temple.edu/~mjastrem
Temple University, Philadelphia )(
$Id: .signature,v 1.4 1996/12/27 21:35:57 mjastrem Exp mjastrem $

User avatar
Thunderstone
Site Admin
Posts: 2504
Joined: Wed Jun 07, 2000 6:20 pm

database integrity while indexing

Post by Thunderstone »




Webinator's database is designed to be usable while walking and indexing.
How is it "not accurate"?
What is your procedure for "indexing your server"?
User avatar
Thunderstone
Site Admin
Posts: 2504
Joined: Wed Jun 07, 2000 6:20 pm

database integrity while indexing

Post by Thunderstone »



At 04:35 PM 1/10/97 EST, you wrote:

I have the same problem. We index with:

NOTE: actual url has been removed....

index_web.sh
============

echo Clearing the database
/usr/local/etc/httpd/htdocs/webinator/bin/gw -wipe

echo Making Database Entries
/usr/local/etc/httpd/htdocs/webinator/bin/gw http://local-url

echo Indexing Entries
/usr/local/etc/httpd/htdocs/webinator/bin/gw -index
============

During walking, all pages show up for any search.

Bob

User avatar
Thunderstone
Site Admin
Posts: 2504
Joined: Wed Jun 07, 2000 6:20 pm

database integrity while indexing

Post by Thunderstone »



..

The default search in Webinator is ranked. A ranked search requires the
index. Non-ranked searches will work without the index.
The database would also be missing a lot of data during the early period
of the rewalk. A better approach for re-indexing your whole site it to walk
and index to a temp database, then switch it to live and delete the old one.
That way users always have a complete database to search against.
Something like:

gw -dnew http://the-url
gw -dnew -index
mv db old
mv new db

Also remember to copy over any top.html or bottom.html you may have
from the old database to the new one. Then remove the old database.