Hi Jeff,

I do not know about Solaris, but what I know is that the memory footprint 
of the indexing increases a lot as the indexing progresses (about 3-4 
times the index size on hp-ux). So if your computer has little memory 
available, there are a large number of processes running, or the total 
memory available is small, then a lot of paging/swapping will be required 
to create the index. The more paging occurs, the slower the process.

Did you check with the command "top" as you index to see how much time is 
consumed, and how much memory is taken? 

Stephane

-----Original Message-----
From: jeffk@europa.bc.ca [mailto:jeffk@europa.bc.ca]
Sent: 27 juillet 1999 03:31
To: swish-e@sunsite.berkeley.edu
Cc: jeffk@europa.bc.ca
Subject: [SWISH-E] swish-e, endless loop on search


I am having a problem with swish-e 1.3.2 on Solaris2.4 (GNU C 2.7.2)
I am creating indexes using the filesystem method.
When I index a large number of files (the index is around 6-8 Mb, which
I understand is not unheard of) the search time becomes endless, the
swish-e executable just hangs.

Generation of the index file works as usual, and I cannot find any
difference in the files from the small group vs. the large group.

Does anyone have any ideas, please respond to my e-mail

Thanks,

Jeff