someone sent me this
this made me laugh and is so good:
- Serverless isn’t.
seems like someone who either doesn’t work in ops full time, or is a bit junior wrote it, but maybe I’m wrong, these ones I took issue with:
- It’s ok to use shell for complex stuff; it often times is easier, faster, and still less of a mess than juggling libraries and dependencies.
gonna have to just go with lol no on this one. Juggling libraries and dependencies usually only happens one time. if that’s your only reason for using shell, you’re probably terrible
- Ok, we all at times keep adding $, {, }, and @ in random places trying to make things work, but still.
i definitely don’t do this because it’s a huge source for some very insidious scripting bugs that are almost impossible to find and debug
- Blocking TCP port 53 traffic leads to very strange failures. Don’t.
this one befuddled me. I know port 53 is typically DNS stuff, that’s about all I know about it, but I can’t think of a reason why anything I work on would need it open, and other than on domain controllers I never leave this one open.
-
- Somewhere in your infrastructure a service you didn’t know uses DNS for endpoint discovery in a very surprising way.
or your entire infrastructure. see this so often
- Do. Not. Monkey. Around. With. /etc/hosts.
lol, fucking what? the only reason this ever matters is if you don’t know what etc hosts actually does, or you make a change in there and then forget you did 3 months later and get confused by it, which is usually resolved by checking if you did something in /etc/hosts you forgot about. I do this so much, I wouldn’t call it a daily thing, but it’s one of the first things I do when testing a new endpoint that I haven’t registered DNS for yet.
- Nobody knows how git works; everybody simply rm -fr && git checkout’s periodically.
not so much an issue I take with this wrt it being true or not, it just astonishes me how common this is, git isn’t complicated or hard to learn.
-
- There are very few network restrictions creative and determined use of ssh(1) port forwarding can’t overcome.
definitely not true
- strace(1)/ktrace(1) doesn’t lie.
unless of course you alias’d them to something else
- Debugging any sufficiently complex open source product is indistinguishable from reverse engineering a black box.|
gotta throw another hard agree here, especially when the model these days seems to be to create an OSS product that will satisfy 95% of production use cases, but make the documentation so fucking awful that you end up buying the enterprise license just for their support package.
- Good software development practices do not always translate well to ops and friends.
this probably needs clarification but I strongly disagree and don’t know how they even came up with this.
- Prod" is just another name for “staging”.
yikes. not anywhere I wanna work.
- Your source of truth lies.
if it did, no way I could maintain continuous employment, I forget what I had for breakfast an hour later
- grep(1) > Splunk (there, I said it)
tell me how you’ve never maintained a production log pipeline without telling me you’ve never maintained a production log pipeline