Tootfinder

Just so that anyone unfamiliar with the fundamentals of #Sysadminnery doesn’t misunderstand:
If OpenAI had wanted a closed testing sandbox, they would have actually constructed and used a closed testing sandbox. THIS IS NOT HARD.
Sure, it's possible that they have shit netadmins and sysadmins who couldn’t contain a test planned by shit devoopsies and project managers unwilling …

I hate when I screw up.
I know, it happens every now and then to everyone. It just sucks so bad.
Make sure your programmatically deployed config files don’t have lingering interface details specific to one machine. Weird things happen. Terrible things.
<sigh>
#Sysadminnery

I swear there is nothing more annoying in my job than the complete and total nightmare of domain names, distinct auth systems, and self-competing branding that fell out of SolarWinds’ dismemberment.
I never saw the value in this mess and still don’t. I only ever touch any of it when it makes noise, which it only does when the tooling itself is breaking in ways that I can't even comprehend based on the alerts.

This was part of my frustration with working in policy enforcement, a.k.a. Abuse Desk. No one bothers to report abuse. That’s a 2-nines “no one.”
The worst embedded autoplaying horse porn email spammers generate at most single-digit-percentage abuse reports. I’m sure social media is much worse, because people have acclimatized themselves to social media being full of garbage that no one will stop.
#Sysadminnery

Main VM handling email and web and DNS died overnight doing its backup, corrupting its virtual disks in the process.
I am so grateful for offline disk repair.
#Sysadminnery

No, having a good backup was NOT the right solution.
It wasn’t ZFS auto-snapshots that saved me.
It wasn’t VMWare auto-snapshots
It wasn’t daily backup of critical config & data to another machine
It was using VMWare Fusion’s ‘sparse' disk format and their recovery tool. On each one of the ‘base' disk segments.
To each OOPS its own recovery tool…
#Sysadminnery