Stories
Slash Boxes
Comments

SoylentNews is people

Meta
posted by NCommander on Sunday November 13 2022, @04:44AM   Printer-friendly
from the that-was-a-lot-of-work dept.

Alright, I've literally been at it for seven hours. Here's what done, here's what needs to done, and if you find something broken, let me know.

  • Upgraded all machines to Ubuntu 22.04 (from 14.04 in most cases)
  • Successfully brought database cluster back up to full service
  • Rebuild web frontend to modern nginx/Ubuntu from very old Gentoo

Here's what's going to happen sometime tonight or tomorrow

  • Outbound SMTP service will be fixed
  • Start fully decommissioning the first set of servers ...

There's likely a lot of things still broken. if you find something broken, leave it in the comments. I'll get to it in the next day. At least we're not on 8 year old software anymore. Current plan after taking care of the above, going to find a good host (probably Fastmail) for editor emails, I do have the backups of everyone inbox should they want it. This has been, very much an exercise in pain. If you want to support me directly: Patreon for recurring, or Ko-fi for one time.

I'll be back in a few hours to work on this more. We raised another $560 USD for charity, and I hope folks had a great time. The next parts I need to just do, so the next stream will likely involve dealing with getting rehash to run in Docker.

~ NCommander

Update - 2022/11/12: DNS resolved, and at least site outbound email should work now

Update - 2022/11/13: Site outbound email is actually working now

 
This discussion was created by NCommander (2) for logged-in users only, but now has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 3, Informative) by janrinok on Sunday November 13 2022, @12:15PM (8 children)

    by janrinok (52) Subscriber Badge on Sunday November 13 2022, @12:15PM (#1279490) Journal

    At the moment the site seems more responsive than it did 24 hours ago - that might just be that there is less traffic blocking the tubes, the French are not on strike (unlikely), or the system is dropping most of my work in a bit bucket somewhere. But it appears currently to be slightly better than it was.

    From the editorial side everything seems to be working as expected but there might still be the odd bug lingering somewhere. Certainly no big stoppers, apart from email which you know about, found so far.

    Starting Score:    1  point
    Moderation   +1  
       Informative=1, Total=1
    Extra 'Informative' Modifier   0  
    Karma-Bonus Modifier   +1  

    Total Score:   3  
  • (Score: 0) by Anonymous Coward on Sunday November 13 2022, @12:23PM

    by Anonymous Coward on Sunday November 13 2022, @12:23PM (#1279492)

    Yes, I also noticed more responsive...and for a moment the green VT100 mode looked a bit sharper too(grin).

    An apt fortune cookie just now?
    > Take everything in stride. Trample anyone who gets in your way.

  • (Score: 2) by RS3 on Sunday November 13 2022, @04:34PM (5 children)

    by RS3 (6367) on Sunday November 13 2022, @04:34PM (#1279528)

    I'm sure most know that software updates usually contain bug fixes, often add security features and layers that actually slow systems down, but sometimes (happily) bring code / algorithm refinements that improve system speed and efficiency.

    Without knowing true technical details, I'd guess the responsiveness improvement is mostly due to mysql cleanup, especially considering it's being referred to as a "cluster". Again, not knowing details, but if it's some kind of mirroring / rsync between 2 or more mysql instances, and one gets out of sync, who/what determines which is the correct one? Something, some process, anyway, and that process will greatly add to latency if it's trying to resolve a discrepancy.

    • (Score: 3, Informative) by NCommander on Sunday November 13 2022, @08:08PM (4 children)

      by NCommander (2) Subscriber Badge <michael@casadevall.pro> on Sunday November 13 2022, @08:08PM (#1279548) Homepage Journal

      No, we literally use Oracle's MySQL Cluster, although its an older community edition. That's on the overhaul list, but at least it's no longer in pieces of colorful confetti.

      --
      Still always moving
      • (Score: 2, Interesting) by fab23 on Sunday November 13 2022, @09:38PM (2 children)

        by fab23 (6605) Subscriber Badge on Sunday November 13 2022, @09:38PM (#1279565) Homepage

        Other options may be MariaDB Galera Cluster or then go to PostgreSQL (this may be a little bit more headache when doing major upgrades).

        • (Score: 2) by RS3 on Monday November 14 2022, @03:25AM (1 child)

          by RS3 (6367) on Monday November 14 2022, @03:25AM (#1279593)

          Is a cluster necessary?

      • (Score: 2) by RS3 on Monday November 14 2022, @03:20AM

        by RS3 (6367) on Monday November 14 2022, @03:20AM (#1279592)

        What do you have against color? :-}

  • (Score: 2) by mrpg on Monday November 14 2022, @12:38AM

    by mrpg (5708) Subscriber Badge <mrpgNO@SPAMsoylentnews.org> on Monday November 14 2022, @12:38AM (#1279579) Homepage

    when I studied French in the book there was an episode about a strike. Years later I went to Paris and guess what, grève des trains. But the train arrived just one hour late, nothing real bad.