dybr
Had a 1-1 with Jacob, pretty short, everything's fine? everything's fine.
While I was working on my shit, tests stopped running, because there was not enough disk space in the postgres machine. Omfg. Wrote to our devops people or something. Karen answered that we can ssh in there and check.
Karen approved my stuff, except for a typo in an exception message, which she suggested to fix; I did. And kaboom, the tests in Jenkins all failed too. I mean, the integration tests.
I was in panic. Ok, I looked through the code.
Problems:
- these integration tests were not running for over a year;
- every test case creates a new db;
- dropping that db was in a wrong branch of the code, so they all stayed.
Fck. Ok, so... what I did. Wait, it was not easy.
I could not even ping the pg db. Why? No vpn. Why? FIIK, my AT&T was blocking it or something.
So I collected via my phone.
Then I had to extract the db password, which was hiding encrypted in a config file. For that, I knew how to do it, run the code in debugger, and voilà. But it did not run, IntelliJ was just hanging, taking 370% of my cores. Had to kill it, restart it, run the debugger, found the password, remembered it forever, and voilà.
So, what I did.
- installed psql (some ancient knowledge)
- found the trick how to do bulk `DROP DATABASE`
- dropped all 308 of those test databases.
Profit. But github did not recognize a rerun in jenkins, so I had to add another empty line in a .sql file, commit it, and profit.
Ok, I have just two days left to finish this functionality.
Oh, and by the way, I'm also on pager duty, answering and confirming all those "limits exceeded" failures.
What I will do. Check that the db deletion works. Maybe merge some test cases into bigger ones (not nice).
And tomorrow I'll have to finish this db preparation part (all tests are already there), because when it comes to prod, I really don't want to log in and start manipulating the prod db manually. No.
So, what do you think. I find it all funny and entertaining.

























