Re: Compression of full-page-writes

On Thu, Jan 8, 2015 at 11:59 PM, Rahila Syed <rahilasyed.90@gmail.com> wrote:
> Below are performance numbers in case of synchronous replication with and
> without fpw compression using latest version of patch(version 14). The patch
> helps improve performance considerably.
> Both master and standby are on the same machine in order to get numbers
> independent of network overhead.
So this test can be used to evaluate how shorter records influence
performance since the master waits for flush confirmation from the
standby, right?

> The compression patch helps to increase tps by 10% . It also helps reduce
> I/O to disk , latency and total runtime for a fixed number of transactions
> as shown below.
> The compression of WAL is quite high around 40%.
>
> Compression                    on
> off
>
> WAL generated             23037180520(~23.04MB)
> 38196743704(~38.20MB)
Isn't that GB and not MB?

> TPS                             264.18                    239.34
>
> Latency average            60.541  ms           66.822
> ms
>
> Latency stddev              126.567 ms           130.434
> ms
>
> Total writes to disk         145045.310 MB     192357.250MB
> Runtime                       15141.0 s  16712.0 s
How many FPWs have been generated and how many dirty buffers have been
flushed for the 3 checkpoints of each test?

Any data about the CPU activity?
-- 
Michael