Horton

Re: Write Ahead Logging for Hash Indexes

Robert Haas <robertmhaas@gmail.com>

View thread View on postgr.es
From: Robert Haas <robertmhaas@gmail.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Stephen Frost <sfrost@snowman.net>, Amit Kapila <amit.kapila16@gmail.com>, Jesper Pedersen <jesper.pedersen@redhat.com>, Jeff Janes <jeff.janes@gmail.com>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2017-03-15T14:51:36Z
Lists: pgsql-hackers 
On Wed, Mar 15, 2017 at 10:34 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> FWIW, I'm not certain that Stephen is correct to claim that we have
> some concrete problem with sparse files.  We certainly don't *depend*
> on sparse storage anyplace else, nor write data in a way that would be
> likely to trigger it; but I'm not aware that we need to work hard to
> avoid it.

That theory seems inconsistent with how mdextend() works.  My
understanding is that we zero-fill the new blocks before populating
them with actual data precisely to avoid running out of disk space due
to deferred allocation at the OS level.  If we don't care about
failures due to deferred allocation at the OS level, we can rip that
logic out and improve the performance of relation extension
considerably.  If we do care about failures due to deferred
allocation, then leaving holes in the file is a bad idea.

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

Commits

  1. Add a regression test for snapshot too old with hash indexes.

  2. hash: Add write-ahead logging support.

  3. Improve coding in _hash_addovflpage.

  4. Remove _hash_wrtbuf() in favor of calling MarkBufferDirty().

  5. Improve hash index bucket split behavior.