Horton

Re: logtape.c stats don't account for unused "prefetched" block numbers

Peter Geoghegan <pg@bowt.ie>

View thread View on postgr.es
From: Peter Geoghegan <pg@bowt.ie>
To: Jeff Davis <pgsql@j-davis.com>
Cc: Alvaro Herrera <alvherre@2ndquadrant.com>, PostgreSQL Hackers <pgsql-hackers@postgresql.org>
Date: 2020-09-14T23:03:14Z
Lists: pgsql-hackers 
On Mon, Sep 14, 2020 at 3:54 PM Peter Geoghegan <pg@bowt.ie> wrote:
> If I add the assertion described above and run the regression tests,
> it fails within "select_distinct" (and at other points). This is the
> specific code:

This variant of the same assertion works fine:

+    Assert(lts->enable_prealloc || lts->nBlocksWritten ==
lts->nBlocksAllocated);

(This is hardly surprising, though.)

-- 
Peter Geoghegan

Commits

  1. Change LogicalTapeSetBlocks() to use nBlocksWritten.

  2. HashAgg: release write buffers sooner by rewinding tape.

  3. logtape.c: do not preallocate for tapes when sorting

  4. Fix bogus MaxAllocSize check in logtape.c.

  5. Avoid fragmentation of logical tapes when writing concurrently.

  6. Logical Tape Set: use min heap for freelist.

  7. Don't create "holes" in BufFiles, in the new logtape code.