Horton

Re: WIP: Avoid creation of the free space map for small tables

Amit Kapila <amit.kapila16@gmail.com>

View thread View on postgr.es
From: Amit Kapila <amit.kapila16@gmail.com>
To: John Naylor <john.naylor@2ndquadrant.com>
Cc: pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2019-02-21T06:58:16Z
Lists: pgsql-hackers

Attachments

On Mon, Feb 11, 2019 at 10:48 PM John Naylor
<john.naylor@2ndquadrant.com> wrote:
>
> On 2/9/19, Amit Kapila <amit.kapila16@gmail.com> wrote:
> > On Tue, Feb 5, 2019 at 3:25 PM John Naylor <john.naylor@2ndquadrant.com>
> > wrote:
> >>
> >> On Tue, Feb 5, 2019 at 4:04 AM Amit Kapila <amit.kapila16@gmail.com>
> >> wrote:
> > This is certainly a good test w.r.t code coverage of new code, but I
> > have few comments:
> > 1. The size of records in test still depends on alignment (MAXALIGN).
> > Though it doesn't seem to be a problematic case, I still suggest we
> > can avoid using records whose size depends on alignment.  If you
> > change the schema as CREATE TABLE fsm_check_size (num1 int, num2 int,
> > str text);, then you can avoid alignment related issues for the
> > records being used in test.
>
> Done.
>

Oops, on again carefully studying the test, I realized my above
comment was wrong.  Let me explain with a test this time:
CREATE TABLE fsm_check_size (num int, str text);
INSERT INTO fsm_check_size SELECT i, rpad('', 1024, 'a') FROM
generate_series(1,3) i;

So here you are inserting 4-byte integer and 1024-bytes variable
length record.  So the tuple length will be tuple_header (24-bytes) +
4-bytes for integer + 4-bytes header for variable length data + 1024
bytes of actual data.  So, the length will be 1056 which is already
MAXALIGN.  I took the new comments added in your latest version of the
patch and added them to the previous version of the patch.   Kindly
see if I have not missed anything while merging the patch-versions?

-- 
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com

Commits

  1. Improve code comments in b0eaa4c51b.

  2. During pg_upgrade, conditionally skip transfer of FSMs.

  3. Add more tests for FSM.

  4. Doc: Update the documentation for FSM behavior for small tables.

  5. Make FSM test portable.

  6. Avoid creation of the free space map for small heap relations, take 2.

  7. Move page initialization from RelationAddExtraBlocks() to use, take 2.

  8. Avoid creation of the free space map for small heap relations.

  9. In bootstrap mode, don't allow the creation of files if they don't already