Hi Andrey,
Thanks again for your tips .. the psync thingy in particular. I need to
verify if that applies to PostgreSQL, because it brings huge gains compared
to sync!
That's easy to explain, it just does one syscall less per IO. It should
indeed bring home a measurable gain as, with synchronous I/O, I believe
you're cpu-limited.
Sadly, it seems PostgreSQL currently does lseek/read/write. (I'll double
check tomorrow running perf against an active PostgreSQL instance).
There was a patch discussed here using pread/pwrite when avail
https://www.postgresql.org/message-id/flat/CABUevEzZ%3DCGdmwSZwW9oNuf4pQZMExk33jcNO7rseqrAgKzj5Q%40mail.gmail.com#CABUevEzZ=CGdmwSZwW9oNuf4pQZMExk33jcNO7rseqrAgKzj5Q@xxxxxxxxxxxxxx
which ends with a comment by Tom Lane (PostgreSQL core developer)
"Well, my point remains that I see little value in messing with
long-established code if you can't demonstrate a benefit that's clearly
above the noise level."
=(
I will post the findings from our discussion here to the PG hackers
list. Maybe ...
Cheers,
/Tobias
--
To unsubscribe from this list: send the line "unsubscribe fio" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
