Oracle vs. PostgreSQL – A Comment • Hacker News Board

Re: Oracle vs. PostgreSQL – A Comment

#201

Post by vbezhenar » Sun, May 31, 2020, 2:17 AM UTC

wbl wrote: post #141

vbezhenar wrote: post #15

That's not a very useful comparison. Databases are made to execute queries, not to install them. The only useful metric is performance and features (and price, of course). Sure, Oracle is bloated and pain to manage, but if it works faster, provides unique features and costs less in the end, people will use it.

Oracle has the De Witt clause. If they had better performance would they have that clause?

One of the reasons for this clause might be the fact that proper benchmarking is hard and wrong benchmarks might hurt sales.

Re: Oracle vs. PostgreSQL – A Comment

#202

Post by msla » Sun, May 31, 2020, 2:27 AM UTC

Justsignedup wrote: post #7

As a developer: - No license fees is huge. My company saves tens of thousands on license fees when we can't really afford it. - The ability to set up environments super quickly is amazing. - Rollback of DDLs is huge for testing migrations - The backup features aren't broken by default. I've spent hours researching how to back up an oracle db and transfer it to another. And at that point versions matter. With postures…

> No license fees is huge. My company saves tens of thousands on license fees when we can't really afford it.

It isn't always about affording it, it's about risk reduction, or the perception thereof:

Every single time the company I work for has to renew a license, it has to once again initiate the long, painful process of finding its ass with both hands. "Why are we paying for this? Who exactly is using it? Why are they using it? How much will this cost? Can we make it cost less? Given that it's impossible for us to negotiate this cost, how can we make the company we're licensing it from not charge us as much money as it's inevitably going to? When is this license due? Why hasn't IT, which has no purchasing authority, fixed this yet?"

With Postgres and CentOS and other OSS, none of that happens. The C-levels never know, so they never get a chance to do their headless-chicken dance routine. Much more relaxing.

Re: Oracle vs. PostgreSQL – A Comment

#203

Post by vbezhenar » Sun, May 31, 2020, 2:30 AM UTC

jojobas wrote: post #73

ethbro wrote: post #66

Earlier quoted context omitted.

Publish benchmarks as an LLC, then just go ahead and declare bankruptcy.

Better yet, publish as a Russian or Chinese resident. Disclose results to a (preferably not Oracle-using) tech newspaper on the condition of anonymity.

Last time I checked, Oracle had some kind of branch in Russia. Not sure whether that clause is enforceable within Russia laws, but they could just bribe police, so they would make one's life quite miserable.

Money is money everywhere. May be even more so in Russia.

I have no idea about China, though.

Re: Oracle vs. PostgreSQL – A Comment

#204

Post by vbezhenar » Sun, May 31, 2020, 2:43 AM UTC

koolba wrote: post #38

vbezhenar wrote: post #15

That's not a very useful comparison. Databases are made to execute queries, not to install them. The only useful metric is performance and features (and price, of course). Sure, Oracle is bloated and pain to manage, but if it works faster, provides unique features and costs less in the end, people will use it.

> That's not a very useful comparison. Databases are made to execute queries, not to install them. The only useful metric is performance and features (and price, of course). Sure, Oracle is bloated and pain to manage, but if it works faster, provides unique features and costs less in the end, people will use it. Initial installation may not matter much but long maintenance windows directly lead to higher costs. If ev…

That depends on a service. For example I'm working on a service which is used by people at working hours (something like 8AM - 19PM, Mon-Fri). I can do anything outside of that window (well, +-30 minutes just to be sure) and nobody's going to pay for it, most likely nobody would even notice.

And for service that's expected to work without interruptions, I think, there should be a way to switch to a secondary database, otherwise those promises are futile. And if there's a way to switch to a secondary database, long patch install time is not a big deal.

Because whether it's 1 minute or 60 minutes, it's interruption nonetheless.

Re: Oracle vs. PostgreSQL – A Comment

#205

Post by dwheeler » Sun, May 31, 2020, 2:45 AM UTC

amenod wrote: post #116

scott_w wrote: post #53

Earlier quoted context omitted.

> I doubt they care about the optics. I’ve never dealt with Oracle myself, yet even I’ve known about their reputation since before I joined the software industry.

Yes, but this looks like they are scared of benchmarks. Weird, right? :)

They aren't scared of benchmarks, because they've rendered benchmarks about their products illegal unless they like the results.

Re: Oracle vs. PostgreSQL – A Comment

#206

Post by vbezhenar » Sun, May 31, 2020, 2:45 AM UTC

ksec wrote: post #100

I am sure there are many areas that Oracle is still technically better than Postgre. What I am interested in, is what are those features / problems. And if Postgre has a roadmap for those improvement. I mean not just Oracle. Outside of HN the majority of dev and cooperation still swears by MySQL. I often wished HN instead of constantly hyping a technology, tell me up front what are the Cons. And hopefully everyone wi…

One reason that I've heard is cluster, Oracle supports clustering while Postgres only provides half-baked solutions. But that was years ago, so I'd like for correction.

Re: Oracle vs. PostgreSQL – A Comment

#207

Post by vbezhenar » Sun, May 31, 2020, 2:48 AM UTC

Someone1234 wrote: post #101

aeyes wrote: post #92

Earlier quoted context omitted.

- query hints - planner hints - better statistics (including ways to manually create statistics) - no vacuum - automatic planner adjustment when true cost was off - better planning for prepared statements - active-active clustering with RAC - vendor support for running on SAN including snapshots - usable query analysis tools provided by the vendor (pg_stat_statements alone doesn't cut it) - datapump much faster than…

Several of these are "features" exist to fix broken parts of Oracle (like query/planner hints, planner adjustment, etc). Officially (per Oracle) the planner should be good enough so you never need to use a hint, but realistically all of these are "features" because the built-in systems aren't good. So, granted, other databases don't have Oracle's hacks to fix their broken stuff but I'm not sure why that is an argumen…

So what do you do when you encounter bug in Postgres planner? File a bug and wait for months or years until fix is released? Hack postgres source and fork it? Using hints sounds like a more sane solution. Nobody's perfect. Being able to work around is better than not being able to.

Re: Oracle vs. PostgreSQL – A Comment

#208

Post by skunkworker » Sun, May 31, 2020, 2:51 AM UTC

Justsignedup wrote: post #7

As a developer: - No license fees is huge. My company saves tens of thousands on license fees when we can't really afford it. - The ability to set up environments super quickly is amazing. - Rollback of DDLs is huge for testing migrations - The backup features aren't broken by default. I've spent hours researching how to back up an oracle db and transfer it to another. And at that point versions matter. With postures…

Having DDL is nice until you are trying to add indexes to a production system and can’t have it locking the entire table. Then you need to add the index concurrently and Postgres does not support concurrent index creation inside of a DDL transaction.

Re: Oracle vs. PostgreSQL – A Comment

#209

Post by dwheeler » Sun, May 31, 2020, 2:57 AM UTC

jojobas wrote: post #45

phoe-krk wrote: post #3

Copying from https://stackoverflow.com/questions/12115397/is-it-against-l... : The standard license ( http://www.oracle.com/technetwork/licenses/standard-license-... ) you agree to when you download software from the Oracle Technology Network (OTN) does state that you're not allowed to disclose benchmarks. You may not: > - disclose results of any program benchmark tests without our prior consent. I sadly suspect that…

Chances are, he lives in a country where essentially saying "by purchasing our product you agree not to tell anyone how bad it is" is unenforceable. That might even include US.

DeWitt clauses (the name for these clauses) have been around since the early 1980s, and have not been struck down yet. I would love to see them disappear, though. For more, see my essay here: https://dwheeler.com/essays/dewitt-clause.html

Re: Oracle vs. PostgreSQL – A Comment

#210

Post by vbezhenar » Sun, May 31, 2020, 3:00 AM UTC

beachy wrote: post #162

pstuart wrote: post #114

Earlier quoted context omitted.

You took time to criticize the post for being upvoted despite not being deep enough / giving enough examples. It's likely upvoted because most of us in the industry have no love for Oracle and confirmation of that bias feels good. Turnabout is fair play: Why don't you give some examples of why a startup doing a greenfield project should use Oracle today?

Data point of 1 - we use Oracle in our greenfield platform (in 2018). My reasoning: - our engineers already know Oracle (this is absolutely reason #1). We don't need to get into religious wars about software, just to write good code quickly. - Oracle is cheap enough for our needs. The price of an oracle RDS instance is immaterial in the grand scheme of things. We are saas so our customers will never be installing our…

I can't say that I saw data loss caused by Oracle. But I saw plenty of cases where Oracle required manual intervention to be able to start up after something gone wrong. I never understood why Oracle does not self heal, because those interventions were obvious. I don't have much production experience with other databases, so I can't comment about that, but it was sometimes as simple as power loss which caused Oracle failing to start until you run some mumbo jumbo from sysdba.