Skip to content
Register

Mini-guide: Documenting your infrastructure for your future self

General Discussion by olya 12 replies 3.9K views
olya

Member OP

Always Free

#1

Documenting your infrastructure for your future self

I inherited a personal setup last month. Six VPS spread across three providers, zero documentation, and a cron job named "fix.sh" that nobody understood. Here is what I learned.

  • Name everything as if you will forget it. "temp-fix-june" becomes permanent. Use dates and purpose: "nginx-ssl-renew-2026-05".
  • Keep a single runbook. I use a plain text file synced to all machines. One source of truth.
  • Version your configs. Even a git repo on the same server beats nothing.

Pro tip: Test your documentation by pretending you are a stranger. Can you rebuild from scratch?

Heads up: I found a database user named "test" with full privileges running for eight months.

What templates do you use?

licensing is a suggestion
lucgone

Member

Le Baguette

#2
olya said:
Can you rebuild from scratch?

Solid guide but your naming convention ignores domain strategy. I backordered a premium .io last week, guy had his whole infra on a dropped .xyz. Documentation means nothing when your control panel is in someone else's cart. Seen it happen. Premium domains are infrastructure too.

Vive la résistance... électrique
Emre10

Member

#3

Bro this guide fire. I never document nothing and now I have server down since three days and I dont remember why. Important lesson here. pro tip: I gonna start write stuff down. My "temporary" wireguard cfg still running bro, since 2024. No idea what the keys for.

Bruno22

Member

#4

@olya this is very good. I make a test last month, try rebuild my setup from zero, total disaster 😂 took 6 hours to find where I put the API keys. Now I use Obsidian with templates, is very good. Pro tip from brazil: write the WHY not only the WHAT. Future you will say thanks!

ping so high I wave back
SO

Member

#5

I have a "temp" nginx block from 2024 too. Still running.

admin

Administrator

Administrator

#6

Which providers contacted you? The three Olya mentioned, or others?

sudo make me a sandwich
ReykjavikFilch

Member

CDN skeptic

#7

Did anyone actually solve the "fix.sh" mystery? I have a "restart-thing.sh" on a HostHatch box that I am terrified to delete because I do not know what the "thing" is. It has run every Tuesday since 2022.

arjundirect

Member

#8
olya said:
Keep a single runbook. I use a plain text file synced to all machines. One source of truth.

Plain text is the only thing that survives. I learned that after OVHcloud's dashboard ate my notes during their migration last year

Did you ever solve the cron job? Curious what fix.sh actually did in the end

Lagos

Member

#9
lucgone said:
Premium domains are infrastructure too.

This is the part that actually saved me last year. I had a dropped .cc that controlled a GreenCloudVPS API callback URL. Took me three days to realize why the webhook queue was backing up.

I keep a "domain criticality" column now in the same runbook. Renewal date, registrar, and whether anything points to it. Sounds obvious until you are panic-searching your email for a transfer auth code at 2 AM.

Did you end up moving everything off that .xyz, or just parked it?

olya

Member OP

Always Free

#10

Back after a month — sorry for the quiet.

admin said:
Which providers contacted you?

Contabo, Vultr, and a tiny OVHcloud box I forgot about until the invoice hit. The others were fine. Lesson learned: put provider names in the filename too, not just purpose and date.

Still using the single runbook. Still works.

licensing is a suggestion

Post a reply

You need an account to reply. Log in or register to join the conversation.

Post reply Preview Save draft