Horton

Re: Write Ahead Logging for Hash Indexes

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Stephen Frost <sfrost@snowman.net>
Cc: Amit Kapila <amit.kapila16@gmail.com>, Robert Haas <robertmhaas@gmail.com>, Jesper Pedersen <jesper.pedersen@redhat.com>, Jeff Janes <jeff.janes@gmail.com>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2017-03-15T14:43:24Z
Lists: pgsql-hackers 
Stephen Frost <sfrost@snowman.net> writes:
> I do see that mdwrite() should handle an out-of-disk-space case, though
> that just makes me wonder what's different here compared to normal
> relations that we don't have an issue with a sparse WAL'd hash index but
> we can't handle it if a normal relation is sparse.

*Any* write has to be prepared to handle errors.  There's always a risk of
EIO, and on a COW filesystem you might well get ENOSPC even when you think
you're overwriting previously-allocated storage.  All that we are doing by
pre-allocating storage is reducing the risks a bit, not guaranteeing that
no error will happen.

			regards, tom lane

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.