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: 2020-09-14T21:24:54Z

Lists: pgsql-hackers

Attachments

0001-LogicalTapeSetBlocks-issue-Tentative-fix.patch (application/octet-stream) patch 0001

On Fri, Sep 11, 2020 at 6:37 PM Peter Geoghegan <pg@bowt.ie> wrote:
> It would be awkward if we just used nBlocksWritten within
> LogicalTapeSetBlocks() in the case where we didn't preallocate (or in
> all cases). Not entirely sure what to do about that just yet.

I guess that that's the logical thing to do, as in the attached patch.

What do you think, Jeff?

-- 
Peter Geoghegan

Commits

Change LogicalTapeSetBlocks() to use nBlocksWritten.
- 6e146a663536 13.0 landed
- c8aeaf3ab31e 14.0 landed
HashAgg: release write buffers sooner by rewinding tape.
- 42a46f5a76ec 13.0 landed
- 3bd35d4f516a 14.0 landed
logtape.c: do not preallocate for tapes when sorting
- 93106d71a18a 13.0 landed
- 075896496394 14.0 landed
Fix bogus MaxAllocSize check in logtape.c.
- 4a4f3bf983b4 13.0 landed
- 0852006a946a 14.0 landed
Avoid fragmentation of logical tapes when writing concurrently.
- 896ddf9b3cd7 13.0 cited
Logical Tape Set: use min heap for freelist.
- c02fdc922301 13.0 cited
Don't create "holes" in BufFiles, in the new logtape code.
- 7ac4a389a7db 10.0 cited