Scott’s Blog

Novell OpenEnterprise Coolsolutions published my paper today about Combining Rsync, Quickfinder and NetStorage for Fun and Profit. All my coolsolutions articles are accessible at Novell from my author page. It’s kindof fun to be able to work out stuff for our own organization in Engineering and then share that stuff with other users to help them out. I also derive loads of benefit in my work from Novell’s support forums and Novell Coolsolutions, so it’s only fair to contribute back. There is some great content on Coolsolutions.

I’ve just completed my installation of a Novell OES Linux Server with Novell Quickfinder. I setup the Quickfinder server to index my own website from within Engineering. It indexed my whole site in about 7 seconds. That kind of makes my whole blog seem kind of small, despite the 188 entries (not counting this one) I’ve made since I started it in August last year.

Anyways, the Quickfinder works pretty well, the crawler is very fast, and it is identical on NetWare and Linux under OES. We are partway through getting the necessary software out in our offices to allow us to have a widespread search system for a schwack of our project related corporate data.

Well, I came to Calgary today to install Novell Search Services on our file server hosting the closed projects of our Calgary office. It was much more hassle than I expected from the piloting in the lab that I had done.
(more…)

I’ve been getting ready to go to Calgary to install the search serveer for the closed project files. The search server in NetWare 6 was created during Novell’s “Orange and Red” period, which is thankfully now over, so the search templates are all bright red and orange in the browser. I’ve modified a custom set in a more subdued blue, so that they go with our Intranet colours a bit better, and added the nice orange mouse-overs that for some reason we like to have. I also stripped the results template to the simplest possible view, removing stuff like showing lots of complicated crap, and printing multiple result pages all at once. The simpler the better for our purposes. I also had to create a custom form to invoke the basic search from our normal Intranet search page. I made up a nice “I Feel Lucky” feaure like the one on Google too.

As I blogged previously, we’ve decided to do a preliminary rollout of our new internal search infrastructure in Calgary, to index and search the closed project data. I’ve been doing pilot installs in the lab to make sure I don’t run into any snags.
(more…)

Preliminary piloting on the Novell Search Engine is finished, and I am sure that it will do what we need it to do. We are able to perform indexing on distributed machines and consolidate our indices to a centralized search server, and push configuration changes out from a primary machine. The search cluster machines also don’t need to be part of the same eDirectory tree, so that works out well for our environment. Now, lacking a formal definition of what exactly the user community wants to be able to search, we’re faced with where to go next.
(more…)

Often times when we do work in our IT lab we are trying to emulate a production network, or at least part of one. The reason for this is obviously that we are trying to test out some modification or enhancement that we want in our production environment. We build a replica of a production server in the lab using images of the boot drives, and tape to restore production data. Most of the time I assume that a quick build of a replica server is all that’s needed, and I can get on with testing my new change or new addition. Sometimes this causal approach to duplicating the production environment bites you in the ass.
(more…)

Now that we’ve got a couple of servers in our lab with legitimate data of the sort that we want to index, we’re starting to learn useful information about using the Novell Search Server in a search cluster in our environment.
(more…)

I’ve got an indexing server running with Netware 6.5 and it’s indexing another server’s data that is made available by simply creating an Apache document root on data directory on the second server. I have 9 GB of data (mostly word processing documents, PDF files, and spreadsheets) and it’s taking about 4.5 hours to index over 10 MBit. It’s also uncompressing any compressed files on my data server’s data directories as it goes. Nice.
(more…)

After buggering around with the Quantum tape drive enough, James grabbed a Compaq drive off a production server, and we are proceeding with the restore to my lab setup for the search server that way.
(more…)