Here's an issue I've been observing (SQL 2014 SP1 CU3):- In a database with 20 tables (clustered key on a bigint identity column, one other index, no constraints/defaults/foreign keys/triggers)- And 20 bulk inserts happening at once, one for each table; batch_size 50000, check_constraints, and keepnulls, from a CSV.- Lots of blocking chains occur where:-- The statements involved are all bulk inserts.-- There is one lead bulk insert, and two or more blocked bulk inserts.-- All are to completely different tables with no relationships with each other.-- The last wait stats sometimes indicate a LOGBUFFER wait on the lead blocker, and PAGELATCH_EX on the blocked processes.Of course I can understand there is a lot of IO pressure but I'm wondering what the blocking resource is. I haven't yet been able to catch it in time to check sys.dm_tran_locks in case it reveals it. I'm using SQL Sentry but it doesn't appear to capture that info in the Blocking SQL tab.
↧