let's be honest, most of us wouldn't last an hour

  • Neckbeard_Prime [they/them,he/him]
    ·
    4 years ago

    More like waiting with baited breath while the most recent backup is restored, and then hoping against hope that the filesystem on the host machine doesn't shit the bed two hours after you give the all-clear. Fun fact: if a web host tells you that your built-to-order dedicated host is good to go and that they initialized the RAID-10 array correctly, don't fucking believe them until you've done a badblocks scan on the RAID volume. (If it's your own hardware, then you probably know better already. And if it's AWS or Azure, I think you may have broken one of the first three or four rules of Fediverse.)

    A previous employer had an ecommerce site running on a LAMP host where this happened -- one of the four drives in the array wasn't initialized correctly for whatever reason, and the server would periodically just crash and restart itself (about every 6 weeks), and there would be data loss on the MariaDB server running on the thing due to filesystem corruption. It must have taken nine months to figure out what in the fuck was happening, especially considering how fast turnover was for techs at the hosting company. I had a lot fewer gray hairs before I had to play Web server admin at that place.

    Godspeed, Lemmy devs. And o7