<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=iso-8859-1">
<META content="MSHTML 6.00.2800.1522" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV>> system<BR>> filePro 5.0.09<BR>> windows 2000 advances
server<BR>> P4 2.4 GHz<BR>> 2 gig of ram<BR><BR><BR>Server also has the
following<BR>Dual processor machine<BR>Ultra ata hard drive (don't know the
exact speed but it's about 36 gigs)<BR>All Terminals Running Remote
Desktop.<BR>I can use any terminal, or even logon from home, and go just as fast
as<BR>being at the server.<BR><BR>Last night I tried the following.<BR>I set
PFBIXBUILD=2 and PFNUMIXBUILD=10000.<BR>Now I built A-P indexes in about 1.5
hrs. I guess thats fast enough, Better<BR>than 15 hrs!<BR><BR><BR>I ran into
another problem with indexing a large file a while back. I have a<BR>menu option
setup like this"/fp/dxmaint - -ra -e". When I try running this<BR>on a very
larger file I get an error about half way through. I think it said<BR>something
about memory. Now that I have reindexed my file, I dont want to<BR>try this to
get the exact error message. When using the menu option above,<BR>filepro seems
like is keeps using memory on top of memory for each index,<BR>until it gets to
the last one. There is something different about manually<BR>selecting each
index and building it. This seems to be the only way I can<BR>build the indexes
on this large file. I don't get it?<BR><BR><BR>Mike
Fedko<BR><BR><BR><BR>><BR>> BTW, I don't want to upgrade to 5.0.14. I am
waiting for next ver.<BR>> Ray said<BR>> that the bug slowing browse
window scrolling and flickering when<BR>> using lots<BR>> of show commands
would be fixed in ver 5.0.15 and above.<BR>><BR>> I have a file with about
three million records. Recently I deleted<BR>> some<BR>> records out of
this file. Now I am getting node errors on a<BR>> particular<BR>> Index. I
rebuilt one index this morning and it took over an hour.<BR>> now add
up<BR>> (a-p times 1 hour). That's 16 hours. There has to be a
faster/better<BR>> way. I<BR>> did try setting PFNUMIXBUF=15000 and that
helped a little. What<BR>> about<BR>> PFNUMIXBUILD.<BR>><BR>> Any
help will be greatly appreciated. I am not looking forward to<BR>>
fifteen<BR>> hours of indexing for one file.<BR>><BR>> Thanks in
Advance, Mike Fedko<BR>><BR>> Mike,<BR>><BR>> Just a "dumb"
question, are you rebuilding these indexes from a<BR>workstation<BR>> or
directly on the server?<BR>><BR>> Most of the responses I have read look
like they are considering a *nix<BR>> system.<BR>><BR>> If you are
building from a workstation, create a local server start up<BR>> script and
rebuild from there. No network contention.<BR>><BR>> Now, if there
is a workstation with a very fast processor and fast drives,<BR>> you might
consider using that machine. But, I would try Ken's suggestion<BR>> and
run it from your server directly.<BR>><BR>> Richard Kreiss<BR>> GCC
Consulting<BR>><BR>><BR>><BR>>
_______________________________________________<BR>> Filepro-list mailing
list<BR>> <A
href="mailto:Filepro-list@lists.celestial.com">Filepro-list@lists.celestial.com</A><BR>>
<A
href="http://mailman.celestial.com/mailman/listinfo/filepro-list">http://mailman.celestial.com/mailman/listinfo/filepro-list</A><BR></DIV></BODY></HTML>