I have got to say right now that the GSA is a dream to work with. Maybe that is a trait of appliances in general -- I am not sure, but the hardest part of getting the GSA up and running was getting it into the rack. I had to take a hacksaw to one of my rack uprights to finally squeeze the GSA home (and apparently I am not the only one!).

The GSA is actually a Dell server running the Google software on top of a hardened version of Linux. The system is completely locked down, which has good and bad points depending upon your point of view. On the downside, it can be very difficult to work out what the appliance is doing when developing any serious integration modules.

The upside, though, is the support. Given the locked down nature of the appliance, Google needs to be able to offer a good support service to their partners and customers and, in my opinion, they offer excellent support. I needed support a couple of times during the integration of the Domino security module and I have found the service to be excellent. All I needed to do was enable SSH access to the appliance and open up the port on our firewall. The Google engineer was then able to log straight into the unit and within a very short space of time, pointed me in the right direction.

Setting up the GSA
Setting up the GSA is simplicity itself. I simply connected the yellow crossover cable to my laptop and the appliance and powered up. After a few minutes, the Network Settings wizard, shown in Figure B, automatically started and after providing some basic network details such as IP address, hostname, and DNS details, the system was ready -- just sitting there whirring away and waiting for content.

FIGURE B


Here, I've completed the Network Settings page on the Google Search Appliance GB-1001. Roll over picture for a larger image.

Configuring the GSA to accept any feed is pretty straightforward. In Figure C, we simply add the URL patterns of the incoming records to the server Crawl URLs screen.

FIGURE C


I'm allowing content to be fed from Domino servers. Roll over picture for a larger image.

This screen is a little misleading as we are not actually asking the GSA to crawl the Domino server at any time, but without an entry in this field, the system will reject all Notes documents submitted in the feed.

The Domino server
With this done, I turned my attention back to the Domino server. I wanted to send some documents from a standard document library database to the GSA, so I registered the database with the connector by completing a database registration document, entered the GSA address into the connector's setup profile, and waited for the magic to happen.

At this point (as every developer knows), the first time we run our software, it all works exactly as it was designed. For me, the following was going to happen without a hitch:

1. The Domino crawler will run.
2. The system will spot the registered database and extract the documents and fields that I told it to.
3. Documents will be split into Notes and attachment records.
4. Document titles and summaries will be generated.
5. Metadata and security tokens will be applied.
6. Documents will be converted to Google XML and fed to the GSA.
7. Stub records of each submitted item will be created for incremental update management.
8. The whole process will repeat without error when the crawler next runs whilst I go to the pub for a well earned beer.