Using Database Engine Tuning Advisor – How to Generate Scripts for Recommendations

From the comments I see you removed the default time limit like I mentioned but still the analysis didn't happen in enough time or ever complete as I feared.

18GB is definitely the largest trace I have ever heard of attempting to be analyzed by the Database Engine Tuning Advisor. I suggest you break your trc file up into smaller sizes. Either by looking at the data a bit more filtered or for a smaller window of time while still maintaining enough of a workload to capture both data modification and selects.

Alternatively -

I haven't been a big fan of the DETA and sometimes the results, even with a properly sized workload file, fly in the face of reality (most index suggestions are very wide covering index in a lot of cases, redundant indexes in overlapping column coverage at times, etc.) I would suggest a "Top 10" style tuning exercise may be a better approach. Find the worst performing queries by duration and/or reads and work on tuning them through query tuning and index tuning/design best practices.

Edited: Added a couple more tips based on comment from OP

I would also look to set an end time. DETA isn't supposed to be a "done in 5 minutes" deal. It is trying to be incredibly thorough and look at your database as deeply as it can and analyze the workload as long as it can to come up with the recommendations. Have you tried setting an end time for the analysis?

I'd check out this link on MSDN for some other tips - https://blogs.technet.com/themes/blogs/generic/post.aspx?WeblogApp=sql_server_isv&y=2011&m=04&d=08&WeblogPostName=fundamentals-running-database-engine-tuning-advisor-and-selecting-indexes&GroupKeys=

Bottom line though - This will be a slower operation. The larger the workload file and more complex the database, the slower it will be. If you are running on the same system that other work is going on then you are each competing for resources.

I expect the analysis is carried out serially, one query at a time, as doing so in parallel would be unreliable.

DTA produces recommendations by creating hypothetical indexes and evaluating the impact on a query's execution plan. If analysis were carried out on multiple queries at a time, the index created for one query could influence the analysis of another.

A hypothetical index is created using the undocumented command CREATE INDEX WITH STATISTICS_ONLY and as the name implies this creates just the statistics for the index, without building the physical structure.