I have a table named ClientContext that has only the clustered index.
I have a procedure sp1 that call other two procedures: sp2 and then sp3.
The procedure sp2 inserts some data in the table and sp3 update data in the table.
CREATE PROCEDURE sp1
CREATE PROCEDURE sp2
INSERT INTO ClientContext.....
CREATE PROCEDURE sp3
The procedure sp1 is called from application. I have an deadlock on UPDATE in sp3 procedure.
I get the graph of deadlock with SQL Profiler (see the attachement Deadlock.png image file).
I get the locks after deadlock ocured from sys.dm_tran_locks system view (see attachement dm_tran_lock.png)
It seems from the graph that process 55 has Update (U) lock on the index and process 64 has exclusive lock (X) on the same index.
But U and X locks are incompatible locks.
How can two processes put incompatible locks on the same resource?
Could anybody explain me this?