Horton

Re: Proposal: Local indexes for partitioned table

Alvaro Herrera <alvherre@alvh.no-ip.org>

View thread View on postgr.es
From: Alvaro Herrera <alvherre@alvh.no-ip.org>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Robert Haas <robertmhaas@gmail.com>, Pg Hackers <pgsql-hackers@postgresql.org>, Maksim Milyutin <milyutinma@gmail.com>, Amit Langote <Langote_Amit_f8@lab.ntt.co.jp>
Date: 2017-11-03T17:30:42Z
Lists: pgsql-hackers 
> Robert Haas <robertmhaas@gmail.com> writes:

> > We could do that, but the motivation for the current system was to
> > avoid leaking memory in a long-lived context.

Yeah, my approach here is to use a CATCH block that deletes the memory
context just created, thus avoiding a long-lived leak.

Tom Lane wrote:

> Another key point is to avoid leaving a corrupted relcache entry behind
> if you fail partway through.

Sure ... in the code as I have it we only assign the local variable to
the relcache entry if everything is succesful.  So no relcache
corruption should result.

> It might work to build the new key in a context that's initially a
> child of CurrentMemoryContext, then reparent it to be a child of
> CacheMemoryContext when done. 

That's another way (than the PG_TRY block), but I think it's more
complicated with no gain.

-- 
Álvaro Herrera                https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

Commits

  1. Local partitioned indexes

  2. Fix StoreCatalogInheritance1 to use 32bit inhseqno

  3. Get rid of copy_partition_key

  4. Simplify index_[constraint_]create API