Horton

Re: speed up a logical replica setup

Amit Kapila <amit.kapila16@gmail.com>

View thread
From: Amit Kapila <amit.kapila16@gmail.com>
To: Euler Taveira <euler@eulerto.com>
Cc: Shlok Kyal <shlok.kyal.oss@gmail.com>, "kuroda.hayato@fujitsu.com" <kuroda.hayato@fujitsu.com>, "pgsql-hackers@lists.postgresql.org" <pgsql-hackers@lists.postgresql.org>, vignesh C <vignesh21@gmail.com>, Michael Paquier <michael@paquier.xyz>, Peter Eisentraut <peter@eisentraut.org>, Andres Freund <andres@anarazel.de>, Ashutosh Bapat <ashutosh.bapat.oss@gmail.com>, Fabrízio de Royes Mello <fabriziomello@gmail.com>
Date: 2024-02-23T04:44:39Z
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 Fri, Feb 23, 2024 at 8:16 AM Euler Taveira <euler@eulerto.com> wrote:
>
> On Wed, Feb 21, 2024, at 5:00 AM, Shlok Kyal wrote:
>
> I found some issues and fixed those issues with top up patches
> v23-0012 and v23-0013
> 1.
> Suppose there is a cascade physical replication node1->node2->node3.
> Now if we run pg_createsubscriber with node1 as primary and node2 as
> standby, pg_createsubscriber will be successful but the connection
> between node2 and node3 will not be retained and log og node3 will
> give error:
> 2024-02-20 12:32:12.340 IST [277664] FATAL:  database system
> identifier differs between the primary and standby
> 2024-02-20 12:32:12.340 IST [277664] DETAIL:  The primary's identifier
> is 7337575856950914038, the standby's identifier is
> 7337575783125171076.
> 2024-02-20 12:32:12.341 IST [277491] LOG:  waiting for WAL to become
> available at 0/3000F10
>
> To fix this I am avoiding pg_createsubscriber to run if the standby
> node is primary to any other server.
> Made the change in v23-0012 patch
>
>
> IIRC we already discussed the cascading replication scenario. Of course,
> breaking a node is not good that's why you proposed v23-0012. However,
> preventing pg_createsubscriber to run if there are standbys attached to it is
> also annoying. If you don't access to these hosts you need to (a) kill
> walsender (very fragile / unstable), (b) start with max_wal_senders = 0 or (3)
> add a firewall rule to prevent that these hosts do not establish a connection
> to the target server. I wouldn't like to include the patch as-is. IMO we need
> at least one message explaining the situation to the user, I mean, add a hint
> message.
>

Yeah, it could be a bit tricky for users to ensure that no follow-on
standby is present but I think it is still better to give an error and
prohibit running pg_createsubscriber than breaking the existing
replication. The possible solution, in this case, is to allow running
pg_basebackup via this tool or otherwise and then let the user use it
to convert to a subscriber. It would be good to keep things simple for
the first version then we can add such options like --force in
subsequent versions.

-- 
With Regards,
Amit Kapila.