> Why does the WT_WITHOUT_SCHEMA_LOCK macro spin & yield when re-acquiring the schema lock, why not just block until it has it?
Alex Gorrod said:
> That is a good question. I copied the logic from WT_WITH_SCHEMA_LOCK, and I believe the same question could be applied there.
> Michael: Is there a reason WT_WITH_SCHEMA_LOCK uses __wt_spin_trylock?
> WT_WITH_SCHEMA_LOCK used to spin so that it could notice if the WT_SESSION_SCHEMA_LOCKED flag was set – this happened during shutdown of LSM worker threads.
> Now that Alex's changes have been merged, spinning should no longer be required.