Comments (11)

  1. […] I covered corruption in nonclustered indexes, the easiest type of corruption to handle. Today, I’m going to move on to something slightly […]

  2. Great series Robert.

    If the corrupt NCI is helping inforce uniqueness are you able to drop the index and then recreate it inside a single transaction? – to prevent possible duplicate values being stored



    1. Good question, Martin. Just tested it, and no problem doing it in a single transaction.

      I also tested to see if you could create a 2nd version of the index with a different name before dropping the old, and that failed as well as it opts to scan the existing index.

  3. Hi Davis,
    great post, hmm, may be you would change the title from “Day 5 of 31 Days of Disaster Recovery: Back That Thang Up” to “Day 5 of 31 Days of Disaster Recovery: Dealing With Corruption in a Nonclustered Index”, ahah.

    1. Thanks Ning! Good catch. I’ve started leaving the title off of the post anyway because it just gets repeated on teh post page.

  4. […] Recovery series. Previously, I’ve talked about several different forms of corruption: Nonclustered Index, Allocation Pages, and Tempdb. these were all fairly simple to fix. Today I’m going to dive […]

Leave a Reply