Re: [Proposal] Fully WAL logged CREATE DATABASE - No Checkpoints

Thank you, Dilip for the quick response. I am okay with the changes done in
the v7 patch.

One last point - If we try to clone a huge database, as expected CREATE
DATABASE emits a lot of WALs, causing a lot of intermediate checkpoints
which seems to be affecting the performance slightly.

--
With Regards,
Ashutosh Sharma.

On Mon, Dec 6, 2021 at 9:59 AM Dilip Kumar <dilipbalaut(at)gmail(dot)com> wrote:

> On Mon, Dec 6, 2021 at 9:17 AM Ashutosh Sharma <ashu(dot)coek88(at)gmail(dot)com>
> wrote:
> >
> > Here are few more review comments:
>
> Thanks for reviewing it.
>
> > 1) It seems that we are not freeing the memory allocated for buf.data in
> CreateDirAndVersionFile().
>
> Yeah this was a problem in v6 but I have fixed in v7, can you check that.
> >
> > + */
> > +static void
> > +CreateDirAndVersionFile(char *dbpath, Oid dbid, Oid tsid, bool isRedo)
> > +{
> >
> > 2) Do we need to pass dbpath here? I mean why not reconstruct it from
> dbid and tsid.
>
> Yeah we can do that but I thought computing dbpath has some cost and
> since the caller already has it why not to pass it.
>
> >
> > 3) Not sure if this point has already been discussed, Will we be able to
> recover the data when wal_level is set to minimal because the following
> condition would be false with this wal level.
> >
> > + use_wal = XLogIsNeeded() &&
> > + (relpersistence == RELPERSISTENCE_PERMANENT || copying_initfork);
> >
>
> Since we are creating new relfilenode this is fine, refer "Skipping
> WAL for New RelFileNode" in src/backend/access/transam/README
>
> --
> Regards,
> Dilip Kumar
> EnterpriseDB: http://www.enterprisedb.com
>