Horton

Re: wake up logical workers after ALTER SUBSCRIPTION

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Nathan Bossart <nathandbossart@gmail.com>
Cc: "houzj.fnst@fujitsu.com" <houzj.fnst@fujitsu.com>, Amit Kapila <amit.kapila16@gmail.com>, Melih Mutlu <m.melihmutlu@gmail.com>, Thomas Munro <thomas.munro@gmail.com>, "Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2023-01-24T18:13:55Z
Lists: pgsql-hackers 
Nathan Bossart <nathandbossart@gmail.com> writes:
> IMO ideally there should be a DSA_HANDLE_INVALID and DSHASH_HANDLE_INVALID
> for use with dsa_handle and dshash_table_handle, respectively.  But your
> patch does seem like an improvement.

Yeah, particularly given that dsa.h says

/*
 * The handle for a dsa_area is currently implemented as the dsm_handle
 * for the first DSM segment backing this dynamic storage area, but client
 * code shouldn't assume that is true.
 */
typedef dsm_handle dsa_handle;

but then provides no way for client code to not be aware that a
dsa_handle is a dsm_handle, if it needs to deal with "invalid" values.
Either that comment needs to be rewritten or we need to invent some
more macros.

I agree that the patch as given is an improvement on what was
committed, but I wonder whether we shouldn't work a little harder
on cleaning this up more widely.

			regards, tom lane

Commits

  1. Avoid type cheats for invalid dsa_handles and dshash_table_handles.

  2. Track logrep apply workers' last start times to avoid useless waits.

  3. Check for two_phase change at end of process_syncing_tables_for_apply.

  4. Wake up a subscription's replication worker processes after DDL.