| From: | Peter Geoghegan <pg(at)bowt(dot)ie> |
|---|---|
| To: | Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> |
| Cc: | Jeff Davis <pgsql(at)j-davis(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: logtape.c stats don't account for unused "prefetched" block numbers |
| Date: | 2020-09-14 22:58:08 |
| Message-ID: | CAH2-Wzn2PhgQNc9k+N3KHz=mRehmR0OmsGYPEsEQcfqDzdNQUQ@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Mon, Sep 14, 2020 at 3:39 PM Peter Geoghegan <pg(at)bowt(dot)ie> wrote:
> I think that they are an exact match in practice (i.e. nBlocksWritten
> == nBlocksAllocated), given when and how we call
> LogicalTapeSetBlocks().
Just to be clear: this is only true for external sorts. The
preallocation stuff can make nBlocksAllocated quite a lot higher.
That's probably why adding a new "Assert(lts->nBlocksWritten ==
lts->nBlocksAllocated)" assertion fails during the regression tests,
though there might be other reasons as well (I'm thinking of the
LogicalTapeRewindForRead() inconsistency Jeff mentioned).
--
Peter Geoghegan
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Peter Geoghegan | 2020-09-14 23:03:14 | Re: logtape.c stats don't account for unused "prefetched" block numbers |
| Previous Message | Ranier Vilela | 2020-09-14 22:57:32 | Re: Simplified version of read_binary_file (src/backend/utils/adt/genfile.c) |