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>, Sawada Masahiko <sawada.mshk@gmail.com>
Cc: Mithun Cy <mithun.cy@enterprisedb.com>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2019-01-30T03:33:32Z
Lists: pgsql-hackers

Attachments

On Tue, Jan 29, 2019 at 8:12 PM John Naylor <john.naylor@2ndquadrant.com> wrote:
>
> On Tue, Jan 29, 2019 at 11:56 AM Amit Kapila <amit.kapila16@gmail.com> wrote:
>
> > You can find this change in attached patch.  Then, I ran the make
> > check in src/bin/pgbench multiple times using test_conc_insert.sh.
> > You can vary the number of times the test should run, if you are not
> > able to reproduce it with this.
> >
> > The attached patch (clear_local_map_if_exists_1.patch) atop the main
> > patch fixes the issue for me.  Kindly verify the same.
>
> I got one failure in 50 runs. With the new patch, I didn't get any
> failures in 300 runs.
>

Thanks for verification.  I have included it in the attached patch and
I have also modified the page.sql test to have enough number of pages
in relation so that FSM will get created irrespective of alignment
boundaries.  Masahiko San, can you verify if this now works for you?

There are two more failures which we need to something about.
1. Make fsm.sql independent of vacuum without much losing on coverage
of newly added code.  John, I guess you have an idea, see if you can
take care of it, otherwise, I will see what I can do for it.
2. I still could not figure out how to verify if the failure on Jacana
will be fixed.  I have posted some theory above and the attached patch
has a solution for it, but I think it would be better if find out some
way to verify the same.

Note - you might see some cosmetic changes in freespace.c due to pgindent.

-- 
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