Horton

Re: speed up a logical replica setup

Euler Taveira <euler@eulerto.com>

View thread
From: "Euler Taveira" <euler@eulerto.com>
To: "Amit Kapila" <amit.kapila16@gmail.com>
Cc: "Shlok Kyal" <shlok.kyal.oss@gmail.com>, "Tomas Vondra" <tomas.vondra@enterprisedb.com>, "kuroda.hayato@fujitsu.com" <kuroda.hayato@fujitsu.com>, "Peter Eisentraut" <peter@eisentraut.org>, "Bharath Rupireddy" <bharath.rupireddyforpostgres@gmail.com>, "pgsql-hackers@lists.postgresql.org" <pgsql-hackers@lists.postgresql.org>, "Michael Paquier" <michael@paquier.xyz>, "Andres Freund" <andres@anarazel.de>, "Ashutosh Bapat" <ashutosh.bapat.oss@gmail.com>, Fabrízio de Royes Mello <fabriziomello@gmail.com>, "vignesh C" <vignesh21@gmail.com>
Date: 2024-05-23T15:13:07Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. pg_createsubscriber: Remove obsolete comment

  2. pg_createsubscriber: Fix an unpredictable recovery wait time.

  3. Fix unstable test in 040_pg_createsubscriber.

  4. Fix the testcase introduced in commit 81d20fbf7a.

  5. Further weaken new pg_createsubscriber test on Windows.

  6. Temporarily(?) weaken new pg_createsubscriber test on Windows.

  7. Make pg_createsubscriber warn if publisher has two-phase commit enabled.

  8. Make pg_createsubscriber more wary about quoting connection parameters.

  9. pg_createsubscriber: Remove failover replication slots on subscriber

  10. pg_createsubscriber: Remove replication slot check on primary

  11. pg_createsubscriber: Only --recovery-timeout controls the end of recovery process

  12. pg_createsubscriber: creates a new logical replica from a standby server

  13. Add some const decorations

  14. Add option force_initdb to PostgreSQL::Test::Cluster:init()

  15. Remove MSVC scripts

On Thu, May 23, 2024, at 5:54 AM, Amit Kapila wrote:
> On Wed, May 22, 2024 at 8:46 PM Euler Taveira <euler@eulerto.com> wrote:
> >
> > Following the same line that simplifies the code, we can: (a) add a loop in
> > check_subscriber() that waits until walreceiver is available on subscriber or
> > (b) use a timeout. The main advantage of (a) is that the primary slot is already
> > available but I'm afraid we need a escape mechanism for the loop (timeout?).
> >
> 
> Sorry, it is not clear to me why we need any additional loop in
> check_subscriber(), aren't we speaking about the problem in
> check_publisher() function?

The idea is to use check_subscriber() to check pg_stat_walreceiver. Once this
view returns a row and primary_slot_name is set on standby, the referred
replication slot name should be active on primary. Hence, the query on
check_publisher() make sure that the referred replication slot is in use on
primary. 

> Why in the first place do we need to ensure that primary_slot_name is
> active on the primary? You mentioned something related to WAL
> retention but I don't know how that is related to this tool's
> functionality. If at all, we are bothered about WAL retention on the
> primary that should be the WAL corresponding to consistent_lsn
> computed by setup_publisher() but this check doesn't seem to ensure
> that.

Maybe it is a lot of checks. I'm afraid there isn't a simple way to get and
make sure the replication slot is used by the physical replication. I mean if
there is primary_slot_name = 'foo' on standby, there is no guarantee that the
replication slot 'foo' exists on primary. The idea is to get the exact
replication slot name used by physical replication to drop it. Once I posted a
patch it should be clear. (Another idea is to relax this check and rely only on
primary_slot_name to drop this replication slot on primary. The replication slot
might not exist and it shouldn't return an error in this case.)


--
Euler Taveira
EDB   https://www.enterprisedb.com/