Estimating MOSS index sizes

Eric Charron has posted about estimating the index size for SharePoint Search.  The formulas are great and provide you a good tool set to walk through the estimating process.  However, I wanted to add the reminder that those formulas are only as effective as the data you use in the beginning.  The first item you need to provide is the size of data to be crawled.  Some time needs to be spent understanding the area you are going to crawl including how much of that data is really going to be indexed.  Without good information to start, all the other calculations will be meaningless and unhelpful.

For example, if you have a 1TB fileshare you are going to index and 3/4 of the share is applications and install files, you can seriously oversize your index space and waste money on unneeded disk space.

UPDATE (May 2008): If you are going to using SharePoint 2007 Service Pack 1, change Eric’s initial size number from 2.5x to 2.85x. They changed the requirement in SP1…see Known Issues: Error: The search request is unable to connect to the search service.

MOSS Search and Hotfixes

In doing a recent Search Proof of Concept for a customer, I ran into a couple errors along the way that needed some fixes and I wanted to document the ones I will be adding to my install documents in the future.  This was a farm environment where there were multiple servers (1 app and 2 wfe) and there were multiple service accounts for each of the different roles within the farm.

 First, when trying to create index schedules, “access is denied” errors were displayed.  MS KB 926959 allowed us to get past the error.

 Second, when trying to view Search Usage reports, we were prompted for credentials even though the account had permissions.  I found the solution in Jingmei Li’s Blog and it points to .Net 2.0 hotfix 913384 and we got past the error once we installed the hotfix.

UPDATE (May 2008): Now, I would recommend you install .Net Framework 2.0 SP1 rather than work to install the individual hotfixes.