I just noticed that my active and hot are 3 days old, it looks like this server is being hit with this bug where the hot calculator dies. Currently, there are fixes in the works and it may be worth cherry picking them or restarting the server during a down time until the fixes make it into the next release.

  • Ada
    shield
    MA
    link
    fedilink
    English
    arrow-up
    19
    ·
    edit-2
    1 year ago

    Unfortunately, we’re stuck with the bug until we upgrade to a version without it. Restarting has been suggested, but the problem with that is that the outbound queue is held in RAM, so it gets lost with a restart.

    We’ll upgrade our version as soon as humanly possible

    • Jon
      link
      fedilink
      English
      arrow-up
      5
      ·
      1 year ago

      Thanks for the update. The outbound queue is held in RAM … what could possibly go wrong? 😂

  • Wander@yiffit.net
    link
    fedilink
    English
    arrow-up
    14
    ·
    edit-2
    1 year ago

    I have set a cronjob on my instance to restart Lemmy via docker_compose every 4 hours. It cause a small 30s downtime every 4 hours, but should be preferable to having the timelines stale.

  • Scurvotron
    link
    fedilink
    English
    arrow-up
    5
    ·
    1 year ago

    Looks like there’s a new release coming soon that hopefully contains a fix for this.

  • HatchetHaro
    link
    fedilink
    English
    arrow-up
    4
    ·
    1 year ago

    Yeah I’ve been wondering about that. Hope this gets fixed soon!