|From:||Fabien COELHO <coelho(at)cri(dot)ensmp(dot)fr>|
|To:||Amit Kapila <amit(dot)kapila16(at)gmail(dot)com>|
|Cc:||Asif Rehman <asifr(dot)rehman(at)gmail(dot)com>, PostgreSQL Developers <pgsql-hackers(at)lists(dot)postgresql(dot)org>|
|Subject:||Re: pgbench - allow to create partitioned tables|
|Views:||Raw Message | Whole Thread | Download mbox | Resend email|
> It is better for a user to write a custom script for such cases.
I kind-of agree, but IMHO this is not for pgbench to decide what is better
for the user and to fail on a script that would not fail.
> Because after that "select-only" or "simple-update" script doesn't
> make any sense. [...].
What make sense in a benchmarking context may not be what you think. For
instance, AFAICR, I already removed benevolent but misplaced guards which
were preventing running scripts without queries: if one wants to look at
pgbench overheads because they are warry that it may be too high, they
really need to be allowed to run such scripts.
This not for us to decide, and as I already said they do if you want to
test no-op overheads. Moreover the problem pre-exists: if the user deletes
the contents of pgbench_accounts these scripts are no-op, and we do not
complain. The no partition attached is just a particular case.
> Having said that, I see your point and don't mind allowing such cases
> until we don't have to write special checks in the code to support such
Indeed, it is also simpler to not care about such issues in the code.
> [...] Now, we can have a detailed comment in printResults to explain why
> we have a different check there as compare to other code paths or change
> other code paths to have a similar check as printResults, but I am not
> convinced of any of those options.
Yep. ISTM that the current version is reasonable.
> [...] I am talking about the call to append_fillfactor in
> createPartitions() function. See, in my version, there are some
Ok, I understand that you want a comment. Patch v15 does that.
|Next Message||Thunder||2019-09-23 07:48:50||Re:PATCH: standby crashed when replay block which truncated in standby but failed to truncate in master node|
|Previous Message||movead li||2019-09-23 04:59:02||Re: row filtering for logical replication|