Here is a simplified version of a table I am having an issue with: create table jobnumbers job_id integer phone_number char(10) connect_time integer I often need to grab all the records for a specific job, so I created a clustered index on job_id. job_id is non-unique. This worked great. I also added a non-clustered index on phone_numbers and job_id together. I have some other code that run's updates like this: update jobnumbers set connect_time = connect_time + x where job_id = y and phone_number = '8885551212' Before I added the clustered index I could do a few thousand of these updates per second. After the clustered index less than 10 per second. It doesn't make any sense to me. I replaced the clustered index with a non-clustered index for the job_id field, by reporting is now much slower but my updating works fine. Anyone seen anything like this and have any suggestions? Thanks in advance.