Re: Can I get some PostgreSQL developer feedback on these five general issues I have with PostgreSQL and its ecosystem?

From: tutiluren(at)tutanota(dot)com
To: Pgsql General <pgsql-general(at)lists(dot)postgresql(dot)org>
Subject: Re: Can I get some PostgreSQL developer feedback on these five general issues I have with PostgreSQL and its ecosystem?
Date: 2020-09-25 05:40:11
Message-ID: MI2lef0--3-2@tutanota.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

> Well not partial as in incremental. Instead dump only some portion of the schema with or without its associated data.
>
It's funny that you should bring that up, considering how it was one of my points... See the point about pg_dump's bug on Windows.

>> I'm saying that PostGIS has a bug due to incorrectly constructed internal queries which makes it impossible to properly name the schema where PostGIS is to reside, causing my database to look very ugly when it has to say "postgis" instead of "PostGIS" for PostGIS's schema. And that was an example of how sloppy/bad third-party things always are, and is one reason why I don't like it when I have to rely on "extensions".
>>
>
> If that is the sum of your issues with PostGIS then I really don't have much sympathy.
>
Why does nobody understand that it was an *example* and not some kind of full PostGIS review?

> They are extensions so you aren't required to use them and rely on their way of doing things. You have the choice of writing your own code/extension or do without completely.
>
It sure is great to have such choices... I can't take it seriously when people say things like this. It's similar to "it's open source so you can easily vet it yourself". It's not taking reality into consideration at all.

As for doing without it, that would make it impossible to deal with GPS coordinates/maps. So it's not really a choice at all.

>> That would entail building an AI into the code that would deal with
>> all the possible OS(versions), Postgres(versions), hardware
>> permutations.
>>
>> I... guess. If "AI" means "a series of ifs". Which is what software... is? I doubt that people who can make the world's most advanced open source database cannot check the amount of RAM and see how fast the CPU/disk is.
>>
>
> It is more then that. It would have to take into account the behavior changes that happen in Postgres between major versions. It also would have to account for OS specific parameters and the changes that happen there between OS versions. It also would need to 'know' how the database was going to be used; readonly, heavy writes, etc. Also how the database should play with other programs on the same machine. Add to the mix containers, cloud instances and so on and you are outrunning the ability of 'ifs' to handle it.
>
If it changes that much, it's far, far worse than I even thought, and it sounds like it will be pointless to even *try* to learn it as it keeps changing between versions/OSes/other stuff.

I can't help but feel as if people just don't want to answer this and other concerns I have. As if there's some silent agreement along the lines of "securing PG DBAs' jobs".

>> Does your server runs to your satisfaction with the default settings?
>>
>> Right now, yes, but that says very little as I'm the only user of it. I've had many nightmares in the past, however, where even determining whether the changes in the config did anything (good or bad) has been impossible. I fundamentally don't like the idea that the config is so "conservative" (crippled) with no obvious/easy way to "set a different general mode". If you honestly think that the numerous performance-related options are easy to understand, I don't know what to say.
>>
>
> The thing is 'general mode' is going to mean something different to someone running a database in the MB-low GB range vs. high GB vs. TB vs. PB.
>
I don't mean this to sound rude, but it's like talking to a wall... What I mean is that there are obviously technical means for software to know whether they are exhausting the system they are running on or not, and expecting people to understand all these intricate internal parameters is just... bizarre. There ought to be some kind of "abstract" setting for those of us who aren't able to (or even *wish* to) comprehend all the PG internals, and just want an efficient database using (roughly) as much of our machine as we want.

This is not the first time I feel like I'm repeating myself over and over in different ways but never getting through. It could be that you are so familiar with PG's internals that it all is obvious to you, but it could just as well be that you don't want to hear about this.

"There's plenty of guides" and "the information is out there" doesn't help me and all the other people who have stuck with the default config and thus a massively restricted PG database for all these years. Just because it's easy to you doesn't mean it's easy to everyone else. Just dealing with composing efficient-enough SQL queries and designing an optimized database structure is (way) more than enough work for most of us. I don't have the luxury of some hired DBA who sits all day tuning the PG server. Besides, I've already explained the privacy issues with that even if I had the money...

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Daniel Westermann (DWE) 2020-09-25 06:51:45 Re: PG 13 trusted extensions and pg_available_extensions
Previous Message tutiluren 2020-09-25 04:49:32 Re: Can I get some PostgreSQL developer feedback on these five general issues I have with PostgreSQL and its ecosystem?