clustered index on ELMAH_Error

[atifaziz]

What new or enhanced feature are you proposing?

Currently, there is no clustered index on ELMAH_Error. From what I understand, this was changed back in r131 as a performance issue. However, this changed the nature of the table to a heap.

I would propose that if ErrorId is not appropriate for a clustered index (it is currently set as the primary key, but non-clustered) then place Sequence as the clustered index with ErrorId remaining as a non-clustered unique index on the table.

Previous suggestion (issue #222) was:

CREATE UNIQUE CLUSTERED INDEX [PK_Sequence] ON [dbo].[ELMAH_Error] 
(
    [Sequence] ASC
)

CREATE UNIQUE NONCLUSTERED INDEX [IX_ErrorId] ON [dbo].[ELMAH_Error] 
(
    [ErrorId] ASC
)

What goal would this enhancement help you achieve?

Better, more consistent performance and adherence with generally recognized best practices.
(see http://msdn.microsoft.com/en-us/library/hh213609.aspx)

---

Originally reported on Google Code with ID 376

Reported by elsimer on 2014-10-26 02:17:36