Re: START_REPLICATION SLOT causing a crash in an assert build - Mailing list pgsql-hackers

From Andres Freund
Subject Re: START_REPLICATION SLOT causing a crash in an assert build
Date
Msg-id 20221006225949.dm3x2nflcaqudzon@awork3.anarazel.de
Whole thread Raw
In response to Re: START_REPLICATION SLOT causing a crash in an assert build  (Kyotaro Horiguchi <horikyota.ntt@gmail.com>)
Responses Re: START_REPLICATION SLOT causing a crash in an assert build  (Masahiko Sawada <sawada.mshk@gmail.com>)
List pgsql-hackers
Hi,

On 2022-10-06 14:10:46 +0900, Kyotaro Horiguchi wrote:
> +1. FWIW, the atttached is an example of what it looks like if we
> avoid file format change.

What about if we go the other direction - simply remove the name from the
stats entry at all. I don't actually think we need it anymore. Unless I am
missing something right now - entirely possible! - the danger that
pgstat_acquire_replslot() mentions doesn't actually exist [anymore]. After a
crash we throw away the old stats data and if a slot is dropped while shut
down, we'll not load the slot data at startup.

The attached is a rough prototype, but should be enough for Jaime to test and
Horiguchi to test / review / think about.

Amit, I CCed you, since you've thought a bunch about the dangers in this area
too.

Greetings,

Andres Freund

Attachment

pgsql-hackers by date:

Previous
From: David Rowley
Date:
Subject: Re: Reducing the chunk header sizes on all memory context types
Next
From: Tom Lane
Date:
Subject: Re: Reducing the chunk header sizes on all memory context types