On 15/07/2019 15:14, Luca Ferrari wrote:
>> Assuming that the 'background activity' writes data, a value of
>> (checkpoint_completion_target) 0.9 means that when your test starts, the
>> system might be still busy in writing data from the previous checkpoint
>> (which started before your pgbench test was launched). That is less likely
>> to happen with a value of 0.1
>
> Uhm...but in the logged table tests a value of 0.9 increases the tps,
> that as far as I understand is in contrast with what you are stating.
What I stated is valid for unlogged tables. (a background checkpoint makes your
pgbench results 'dirty')
When you talk about logged tables, you actually want to spread the checkpoint
over time. The more it is spread, the better performances. But here, probably,
checkpoint has a lot to write compared to the data produced by background job
(and, maybe, checkpoints are happening more frequently?).
>
> Anyway, I'll test more and report back some more results.
good, let us know and do not forget to provide the log lines produced by the
checkpoints too.
regards,
fabio pardi
