We decided that we needed to crawl the index server and not the web servers in our sharepoint environment. We are also using F5 for load balancing the web servers. There are some notes and articles that describe methods of setting up the crawl on a specific web server from the web front end servers but not one that specifically uses the index server. This method by the way can be done on an additional server that is not the index server and is not part of the F5 web balanced servers. I think the previous sentence would be perfect for those environments where they can afford the extra server.
So how do we do this?
We encapsulated both the alternate access mappings for the F5 load balancing and the server name mapping for the index server crawl. Then set the configure office sharepoint server search service setting to not use a dedicated front end, that keeps the host file from being overwritten.
The normal approach would be to set the configure office sharepoint server search service setting to not use a dedicated front end, that keeps the host file being overwritten, in our case the F5 overruled the settings we would put in the host file and get lost on the index server. This is where we used the server name mapping.
By setting the server name mapping to index on the index server and to reflect the F5 alias we were able to force the crawl to the index server but the users would then hit the F5 alias and be load balanced in their search.
We found ourselves in situation where if we tried to use a dedicated index server that was not part of the F5 alias that the index server would get lost and the crawl would fail. It would reflect an error that indicated it could not see the server, even after we set the ip’s in the host file to be the index crawler.
So here is how we made it work.
For the alternate access mappings we created an Internal url pointing to http://F5alias.com with the zone set to default and a public url pointing to https://f5alias.com .
Then an additional alternate access mapping that set the internal url pointing to http://indexserver.com with the zone set to custom and a public url of http://indexserver.com .
Then we created a server name mapping for an indexing address of http://indexserver.com and address for display in search results of https://F5alias.com .
When the crawl is processed it creates the index on the indexserver and displays the f5 load balance alias. The user hits the f5 load balance alias and is pointed to one of the servers in the f5.
