Re: Slow standby snapshot - Mailing list pgsql-hackers

Hello, Simon.

Sorry for calling you directly, but you know the subject better than
anyone else. It is related to your work from 2010 - replacing
KnownAssignedXidsHash with the KnownAssignedXids array.

I have added additional optimization to the data structure you
implemented. Initially, it was caused by the huge usage of CPU (70%+)
for KnownAssignedXidsGetAndSetXmin in case of long (few seconds)
transactions on primary and high (few thousands) max_connections in
Postgres 11.

After snapshot scalability optimization by Andres Freund (2), it is
not so crucial but still provides a significant performance impact
(+10% TPS) for a typical workload, see benchmark (3).

Last patch version is here - (4).

Does such optimisation look worth committing?

Thanks in advance,
Michail.

[1]:
https://github.com/postgres/postgres/commit/2871b4618af1acc85665eec0912c48f8341504c4#diff-8879f0173be303070ab7931db7c757c96796d84402640b9e386a4150ed97b179
[2]: https://commitfest.postgresql.org/29/2500/
[3]:
https://www.postgresql.org/message-id/flat/CANtu0ohzBFTYwdLtcanWo4%2B794WWUi7LY2rnbHyorJdE8_ZnGg%40mail.gmail.com#379c1be7b8134ada5a574078d51b64c6
[4]:
https://www.postgresql.org/message-id/flat/CANtu0ogzo4MsR7My9%2BNhu3to5%3Dy7G9zSzUbxfWYOn9W5FfHjTA%40mail.gmail.com#341a3c3b033f69b260120b3173a66382