Physical - Not VM

Model : ProLiant DL580 G7

CPU : 4x10

Memory : 512 GB

OS : Windows 2008R2 ENT

Max Server Memory = 20 GB

Since posting on this site, I increased the memory allocation for this SQL instance to 40 GB. This eliminated the FAIL ALLOCATION event, and prevented SQL from hanging. Tracing the CACHESTORE_PHDR clerk yielded something interesting. It briefly grows from 38 MB to 17GB and then shrinks to less than 1 MB.

Timestamp size_kb

================ =======

2015-08-31 02:20:45.651 328

2015-08-31 02:20:30.649 17075664

2015-08-31 02:20:15.648 12631408

2015-08-31 02:20:00.642 38472

PHDR grows likes this 4-times per day at the same times each day. The following message appears in the SQL log each time PHDR grows large:

"The query processor ran out of internal resources and could not produce a query plan. This is a rare event and only expected for extremely complex queries or queries that reference a very large number of tables or partitions..."

Questions:

How may I identify the offending query?

Is this expected SQL behavior?