Softaculous creates more problems than it solves

Docker compose is still too much GUI for my blood.

I run everything on a $3 RackNerd in Los Angeles, plain LXC, ansible-pull from git. WordPress core updates via wp-cli cron. No Softaculous, no panel, no problem. If container breaks, ansible rebuilds in 90 seconds.

Friend ask me "where is the button" I say "there is no button, go away"

Osaka represent though. @ronwit we should drink and complain about cheap VPS together.

This is the energy I need (´・ω・｀)

But my friend shop, they need to change menu price at 2am when drunk. They cannot ansible-pull. They need button. This is why I am trapped in docker compose middle ground, forever explaining volumes to crying people at midnight.

Osaka drink yes. I know place in Namba, cheap whiskey, we can cry about our uptime.

I am not understand why anyone run database-backed CMS on cheapest possible VPS.

My PostgreSQL instances live on proper Vultr cloud in Helsinki region, replicated to Stockholm. The wordpress people, they put MySQL on same $2.50 machine as apache, then surprised when OOM killer come.

Softaculous is not the disease. The disease is "I want enterprise result with potato budget and zero knowledge." Softaculous is just symptom that make the disease faster.

Finally someone say it.

I run Gentoo at home, compile everything, I know exactly what is on my system. But I also manage Arch server for local hackerspace — https://wiki.archlinux.org — and we use Softaculous there because other members are not Gentoo people, they are normal humans who want to blog about their cat.

The elitism in this thread is thick enough to spread on bread. Not everyone has time to be you. Some people have jobs, children, lives outside of terminal.

Softaculous is fine if you accept what it is: training wheels. Problem is adults who never take training wheels off, then complain bicycle is slow.

Yes ạ this is exactly my client ạ

They want cheap CDN for game community but also want wordpress blog with Softaculous ạ I say "pick one, your VPS from RackNerd is $4" they say "why cannot both" ạ

Now I am separate them ạ static CDN on RackNerd Los Angeles, wordpress on Hostinger Singapore ạ more expensive but at least when blog kena defaced the CDN still serve game patch ạ

Compartmentalization, not elimination ạ

This. This right here.

I used to run everything on one Vultr instance like a genius. Email, web, git, the works. One defacement, total rebuild, lost everything. Now I spread like Chen's five providers except I am poor so it is three cheap ones.

Softaculous lives in its own jail now. If it burns, it burns alone.

This is not compartmentalization, this is fragmentation with extra steps.

If you cannot describe your entire infra in one terraform plan, you do not understand your infra. "Spread across providers" just means "more dashboards to check when things break at 3am."

I have seen this movie. The sequel is always "why did the backup fail" and the answer is "because backup was on the third cheap provider I forgot about."

Be careful.

You are not wrong but also you are very Estonian about this lor

My five providers: Vultr for compute, Hostinger for cheap shared experiment, another for DNS, another for email, another for... I forget actually. Yes this is the problem.

But the point is not perfect architecture. The point is "my friend cannot blame me when his Softaculous wordpress kena hack because it is not on my Vultr account." Social compartmentalization also important.

Following this thread because I am exactly the person Chen describe. Softaculous wordpress 2022, never update, still running. Should I update now or is it too late, will it break

Also what is terraform

Update now. Backup first. Hostinger has backup button in panel, use it.

Then after update works, set calendar reminder. Every month. Do not be Pierre who forgets 30% of time.

Terraform is... how to say... very advanced Pierre. You do not need terraform. You need discipline.