Re: Few comments on commit 857f9c36 (skip full index scans )

Hi,

Review comments on commit 857f9c36:
1.
@@ -2049,6 +2055,10 @@ _bt_newroot(Relation rel, Buffer lbuf, Buffer rbuf)
metapg = BufferGetPage(metabuf);
metad = BTPageGetMeta(metapg);

+ /* upgrade metapage if needed */
+ if (metad->btm_version < BTREE_VERSION)
+ _bt_upgrademetapage(metapg);
+

The metapage upgrade should be performed under critical section.

2.
@@ -191,6 +304,10 @@ _bt_getroot(Relation rel, int access)
LockBuffer(metabuf, BUFFER_LOCK_UNLOCK);
LockBuffer(metabuf, BT_WRITE);

+ /* upgrade metapage if needed */
+ if (metad->btm_version < BTREE_VERSION)
+ _bt_upgrademetapage(metapg);
+

Same as above.

In other cases like _bt_insertonpg, the upgrade is done inside the
critical section. It seems the above cases are missed.

3.
+ TransactionId btm_oldest_btpo_xact; /* oldest btpo_xact among of
+ * deleted pages */

/among of/among all

Attached patch to fix the above comments.

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

Hi!

On Mon, May 28, 2018 at 11:52 AM, Amit Kapila <amit(dot)kapila16(at)gmail(dot)com>
wrote:

> Review comments on commit 857f9c36:
> 1.
> @@ -2049,6 +2055,10 @@ _bt_newroot(Relation rel, Buffer lbuf, Buffer rbuf)
> metapg = BufferGetPage(metabuf);
> metad = BTPageGetMeta(metapg);
>
> + /* upgrade metapage if needed */
> + if (metad->btm_version < BTREE_VERSION)
> + _bt_upgrademetapage(metapg);
> +
>
> The metapage upgrade should be performed under critical section.
>
> 2.
> @@ -191,6 +304,10 @@ _bt_getroot(Relation rel, int access)
> LockBuffer(metabuf, BUFFER_LOCK_UNLOCK);
> LockBuffer(metabuf, BT_WRITE);
>
> + /* upgrade metapage if needed */
> + if (metad->btm_version < BTREE_VERSION)
> + _bt_upgrademetapage(metapg);
> +
>
> Same as above.
>
> In other cases like _bt_insertonpg, the upgrade is done inside the
> critical section. It seems the above cases are missed.
>
> 3.
> + TransactionId btm_oldest_btpo_xact; /* oldest btpo_xact among of
> + * deleted pages */
>
> /among of/among all
>
> Attached patch to fix the above comments.
>

Thank you for catching this!
Your patch looks good for me.

------
Alexander Korotkov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company

On Mon, May 28, 2018 at 4:52 AM, Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> wrote:
> Hi,
>
> Review comments on commit 857f9c36:
> 1.
> @@ -2049,6 +2055,10 @@ _bt_newroot(Relation rel, Buffer lbuf, Buffer rbuf)
> metapg = BufferGetPage(metabuf);
> metad = BTPageGetMeta(metapg);
>
> + /* upgrade metapage if needed */
> + if (metad->btm_version < BTREE_VERSION)
> + _bt_upgrademetapage(metapg);
> +
>
> The metapage upgrade should be performed under critical section.
>
> 2.
> @@ -191,6 +304,10 @@ _bt_getroot(Relation rel, int access)
> LockBuffer(metabuf, BUFFER_LOCK_UNLOCK);
> LockBuffer(metabuf, BT_WRITE);
>
> + /* upgrade metapage if needed */
> + if (metad->btm_version < BTREE_VERSION)
> + _bt_upgrademetapage(metapg);
> +
>
> Same as above.
>
> In other cases like _bt_insertonpg, the upgrade is done inside the
> critical section. It seems the above cases are missed.
>
> 3.
> + TransactionId btm_oldest_btpo_xact; /* oldest btpo_xact among of
> + * deleted pages */
>
> /among of/among all
>
> Attached patch to fix the above comments.
>

Thank you for reviewing and the patch. All changes looks good to me.

Regards,

--
Masahiko Sawada
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center

> The metapage upgrade should be performed under critical section.

Agree. But after close look I found that btm_version change isn't wal-logged
(near line 2251 in _bt_newroot). So btm_version is not propagated to
replica/backup/etc.
I believe it should be fixed.

--
Teodor Sigaev E-mail: teodor(at)sigaev(dot)ru
WWW: http://www.sigaev.ru/

On Wed, May 30, 2018 at 9:09 PM, Teodor Sigaev <teodor(at)sigaev(dot)ru> wrote:
>> The metapage upgrade should be performed under critical section.
>
>
> Agree. But after close look I found that btm_version change isn't wal-logged
> (near line 2251 in _bt_newroot). So btm_version is not propagated to
> replica/backup/etc.
>

I think it will always be set to BTREE_VERSION (See _bt_restore_meta).
I see that other places like _bt_insertonpg,
_bt_update_meta_cleanup_info, etc. that log meta page contents don't
log version number, so if we want to log it, all such places also need
to be modified, but I don't see the need for same.

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

> I think it will always be set to BTREE_VERSION (See _bt_restore_meta).
You are right, pushed. Thank you!

--
Teodor Sigaev E-mail: teodor(at)sigaev(dot)ru
WWW: http://www.sigaev.ru/