• 0 Posts
  • 23 Comments
Joined 2 years ago
cake
Cake day: January 16th, 2024

help-circle




  • Correct, though to be pedantic anyone can be a CA- you just generate a cert with the right bits to say it’s a ca certificate and then use it to sign any other certificate you want.

    But the only devices that will consider your signature worth anything are ones you also install your ca certificate on. So it’s useful and common in internal networks but isn’t really what is being asked here.

    The hard part is getting in the root CA store of operating systems and browsers. As far as I know they are all maintained independently with their own requirements.









  • growing it like a garden is a perfect phrase imo

    because on windows or Mac it may have just worked. …until it doesn’t, or leaves your windows scaled wrong or placed on monitors that don’t exist or some other failure condition. at which point you reboot and hope for the best.

    when it doesn’t work on Linux I’d check logs, actual configuration, and even the source if I need to.and then I’d hopefully improve things and make it work the way I want it to.


  • If adopt systems then the question is easy to answer: no, journald does everything you need.

    without adopting systemd… well. Are you evaluating going without any log handling at all and maybe just dumping logs ephemerally to tty0? DIYing all log stuff like your init scripts DIY things?

    Personally if I had to go without journald I’d probably go back to using syslog-ng. But I guess there’s an argument for shipping straight into something like opentelemetry-collector if you’re willing to put in a lot of work.




  • not OP but I’d love something like this for a few reasons.

    Sometimes I’m debugging really complicated things and it gets hard to keep track of the info I’ve captured and what I’ve learned, and sometimes you want to recheck some earlier assumption or you learn something new and want to look through older data captured to see if it aligns with newer understandings

    Or it’s a long standing thing and need to step away and come back and refresh your memory of the current understanding. And especially when others might also be working on the same problem and you want to collaborate better.

    Though I am SRE and thinking of debugging issues in overall systems spanning multiple codebases, hosts, and networks. not just specific bugs in a single codebase like I think OP is doing. So I’m also curious if any tool would actually fit both use cases or if being perfect for one would make it not useful for the other.