EDIT:

The pictrs migration is still taking place in the background, and there might be some problems displaying new or some old images, as well as a bit slower site performance. But it seems like the site can run well now.


The lemmy backend makes use of pict-rs for hosting images. A new version (0.5.0) was recently released, and upgrading from the previous version involves a migration that is expected to take a few hours.

I am planning to run that migration either tomorrow evening or on Saturday. If all goes well, the instance will be down for only a few hours.

Just a heads up in case you stop by and see that the site is unavailable.

  • Shdwdrgn
    link
    fedilink
    English
    arrow-up
    3
    ·
    11 months ago

    Does this have anything to do with the issues earlier today (maybe 9-10 hours ago)? After several failed attempts to open new messages or reload the page, I got a message stating that the server was rebooting, and a couple minutes later everything was working fine again.

    • SalamanderOPMA
      link
      fedilink
      arrow-up
      2
      ·
      11 months ago

      That’s probably when I ran a quick upgrade from 0.19.2-rc-5 to 0.1.9.2! If I know it will take only ~5 mins I usually just go ahead without notice.

      • Shdwdrgn
        link
        fedilink
        English
        arrow-up
        2
        ·
        11 months ago

        … and just hope nobody else notices? 😆 I figured something was going on and it would be back shortly, but I also know sometimes servers do thing that admins weren’t aware of, figured I would just check in.

        • SalamanderOPMA
          link
          fedilink
          arrow-up
          2
          ·
          11 months ago

          Yeah, since it is such a short downtime - comparable to what one might randomly see anyway when the CPU spikes for a mysterious reason, I figure that it is not so helpful to plan ahead and make an announcement. I just do it when I have some time.