4 ms·

Thanks for the insights! I'm actually using MySQL (MariaDB, actually) in production because it does work fine, does all we need in an OLTP data store, ops tool

by lima 10y ago

Thanks for the insights! I'm actually using MySQL (MariaDB, actually) in production because it does work fine, does all we need in an OLTP data store, ops tooling is very good and we need reliable multi-master (Galera) for load balancing. I'm not an expert on databases, but I've heard so many DBAs tell me how PostgreSQL is superior with regards to features + correctness (features that I don't actually need for our use case, but still) except for scaling out. What's your take on this? If you were to start on a green field, what database would you recommend for a typical SaaS/OLTP workload?

evanelias 10y ago
I'd say it depends on your team's level of knowledge on various DB's, the expertise available in your geographic area, and the expected size that your data set would grow to. I'm not sure there is a "typical" workload in the general case ;) Don't get me wrong; Postgres is a really really good piece of software with many appealing qualities. But its fanbase is a bit overly rabid, or at least over-eager to trash MySQL for things that haven't been true in years or never were true in the first place. There are some very valid MySQL complaints mixed in there too, but it can take some effort to separate them out...
- lima 10y ago
  Thanks! Our team is definitely going to stick with MySQL (no reason not to) and I have no reason to hate it, but I always like these sort of comparisons since you learn a lot about the technology. I'm really interested in PostgreSQL for "real time OLAP" use cases with things like Citus - apparently, many companies use Citus in particular to run distributed analytics queries on their main database.
- i336_ 10y ago
  Hmm. Have you seen any wholesale feature comparisons between the two that you'd consider objective and fair, and which debunk old myths in the process?
  - evanelias 10y ago
    Unfortunately I haven't. People who are equally well-versed on both databases seem to be the rarest of unicorns.

by lima 10y ago

evanelias 10y ago

I'd say it depends on your team's level of knowledge on various DB's, the expertise available in your geographic area, and the expected size that your data set would grow to. I'm not sure there is a "typical" workload in the general case ;) Don't get me wrong; Postgres is a really really good piece of software with many appealing qualities. But its fanbase is a bit overly rabid, or at least over-eager to trash MySQL for things that haven't been true in years or never were true in the first place. There are some very valid MySQL complaints mixed in there too, but it can take some effort to separate them out...

lima 10y ago
Thanks! Our team is definitely going to stick with MySQL (no reason not to) and I have no reason to hate it, but I always like these sort of comparisons since you learn a lot about the technology. I'm really interested in PostgreSQL for "real time OLAP" use cases with things like Citus - apparently, many companies use Citus in particular to run distributed analytics queries on their main database.
i336_ 10y ago
Hmm. Have you seen any wholesale feature comparisons between the two that you'd consider objective and fair, and which debunk old myths in the process?
- evanelias 10y ago
  Unfortunately I haven't. People who are equally well-versed on both databases seem to be the rarest of unicorns.