potiuk commented on pull request #18506: URL: https://github.com/apache/airflow/pull/18506#issuecomment-975634568
> Is the MariaDB support decision finalized yet? I don't think it is - I think if you want to get support, the best way to make the decision is to state your position as someone who wants it in the discussion in the devlist and argue for it. I think I was the only one (from those speaking so far) who thinks that we should add tests and full support. For me the current status quo is "we do not support MariaDB". Full stop. Even though we know there are many of our users who use MariaDB, they do it fully on their own - and I am going to mostly close/convert into discussion any issues where MariaDB is mentioned and it will look like "maybe a a database issue" (this is very much consequences of the status quo). At least that's my understanding of it (though I think it woudl be better for our users to support it). And just to elaborate a little why I take this 'stance': I believe eiether we say "we support mariadb" - and add tests, or we say "we do not support it" when we have no tests. I think there is no middle-ground and "best-effort" when we (as community) do "no effort" to make sure the support is there, and especially no effort to make sure it is not broken in the future (this is what tests are really about). We are adding new features and there was at least one case where lack of certain DB feature caused change of direction of the implementation (https://cwiki.apache.org/confluence/display/AIRFLOW/AIP-42%3A+Dynamic+Task+Mapping) - big part of the implementation was changed because MySQL does not support big indexes. And without tests we have no guarantee that we will not hit similar limitations of MariaDB in specific versions, especialy that MariaDB versioning is very different from MySQL and some features are appearing with various versions and different cadence. So currently we are not even able to say which version of MySQl our users could use. So for me at least status quo is "we have no tests" which is equivalent to "we do not support it". I was not able to gain enough support for "supporting" MariaDB, so without more support and discussion on the devlist, the status quo is "we do not support MariaDB". But I think if you could chime in from the user perspective and say how much you would like - and why - have support for MariaDB on the devlist, might encourage others to speak and possibly support it or at least revive the discussion. Maybe if we know how many of our users are actually using MariaDB could help (but I think for that we would need to wait for our annual survey - I am definitely planning to add such question for "MySQL" users) - the case would be stronger, but I have not enough data to back my "feeling" that we have a lot of users using MariaDB - possibly even more than MySQL users. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
