Horton

Re: Skipping logical replication transactions on subscriber side

Masahiko Sawada <sawada.mshk@gmail.com>

View thread View on postgr.es
From: Masahiko Sawada <sawada.mshk@gmail.com>
To: Peter Eisentraut <peter.eisentraut@enterprisedb.com>
Cc: Amit Kapila <amit.kapila16@gmail.com>, "houzj.fnst@fujitsu.com" <houzj.fnst@fujitsu.com>, "tanghy.fnst@fujitsu.com" <tanghy.fnst@fujitsu.com>, "osumi.takamichi@fujitsu.com" <osumi.takamichi@fujitsu.com>, Alexey Lesovsky <lesovsky@gmail.com>, PostgreSQL-development <pgsql-hackers@postgresql.org>, Greg Nancarrow <gregn4422@gmail.com>
Date: 2021-10-01T01:00:05Z
Lists: pgsql-hackers 
On Fri, Oct 1, 2021 at 5:05 AM Peter Eisentraut
<peter.eisentraut@enterprisedb.com> wrote:
>
> On 30.09.21 07:45, Masahiko Sawada wrote:
> > I've attached updated patches that incorporate all comments I got so
> > far. Please review them.
>
> I'm uneasy about the way the xids-to-be-skipped are presented as
> subscriptions options, similar to settings such as "binary".  I see how
> that is convenient, but it's not really the same thing, in how you use
> it, is it?  Even if we share some details internally, I feel that there
> should be a separate syntax somehow.

Since I was thinking that ALTER SUBSCRIPTION ... SET is used to alter
parameters originally set by CREATE SUBSCRIPTION, in the first several
version patches it added a separate syntax for this feature like ALTER
SUBSCRIPTION ... SET SKIP TRANSACTION xxx. But Amit was concerned
about an additional syntax and consistency with disable_on_error[1]
which is proposed by Mark Diliger[2], so I’ve changed it to a
subscription option. I tried to find a policy of that by checking the
existing syntaxes but I could not find, and interestingly when it
comes to ALTER SUBSCRIPTION syntax, we support both ENABLE/DISABLE
syntax and SET (enabled = on/off) option.

> Also, what happens when you forget to reset the xid after it has passed?
>   Will it get skipped again after wraparound?

Yes. Currently it's a user's responsibility. We thoroughly documented
the risk of this feature and thus it should be used as a last resort
since it may easily make the subscriber inconsistent, especially if a
user specifies the wrong transaction ID.

Regards,

[1] https://www.postgresql.org/message-id/CAA4eK1LjrU8x%2Bx%3DbFazVD10pgOVy0PEE8mpz3nQhDG%2BmmU8ivQ%40mail.gmail.com
[2] https://www.postgresql.org/message-id/DB35438F-9356-4841-89A0-412709EBD3AB%40enterprisedb.com

-- 
Masahiko Sawada
EDB:  https://www.enterprisedb.com/

Commits

  1. Test ALIGNOF_DOUBLE==4 compatibility under ALIGNOF_DOUBLE==8.

  2. Reorder subskiplsn in pg_subscription to avoid alignment issues.

  3. Add ALTER SUBSCRIPTION ... SKIP.

  4. Optionally disable subscriptions on error.

  5. Update docs of logical replication for commit 8d74fc96db.

  6. Respect permissions within logical replication.

  7. Fix regression test failure caused by commit 8d74fc96db.

  8. Add a view to show the stats of subscription workers.

  9. Add logical change details to logical replication worker errcontext.

  10. Rename LOGICAL_REP_MSG_STREAM_END to LOGICAL_REP_MSG_STREAM_STOP.

  11. Fix typo in protocol.sgml.

  12. Remove unused argument in apply_handle_commit_internal().

  13. Fix replication of in-progress transactions in tablesync worker.

  14. Reorder pg_sequence columns to avoid alignment issue