> I do what you are describing for fun, but I would never recommend it for a business, even if it is reliable.
I appreciate your response and don't necessarily disagree. Could you tell me a little more about why you wouldn't recommend it for a business, even if it is reliable?
> To get around the reliability problems, I have the on-prem laptop environment update cloudflare entries with keep-alive timestamp messages. .. I then have a Google script that monitors the keepalive and if too much time has passed, a “failover” is done and everything that was running on-prem is spun up in digital ocean. The failover script completes by updating DNS and pointing to DO
This is a clever solution. Thank you for your thoughts!
Yes, I wouldn’t recommend it for a business just due to the complexity of the setup and the necessary knowledge to keep it running.
If you’re able to document the setup properly and have multiple people that have intimate knowledge of your machinations.. then you are covered .. unfortunately most people do not cover these bases.
If you have built yourself into this process you will realize quickly that nobody else understands it, and you will have a hard time taking vacation (among other things)
Also you put the business at risk because you need to plan for what the company will do when you get hit by a bus.
For these reasons it’s generally more preferable to pay a premium to go with a standard solution as opposed to something you cobbled together to save small amounts of money.
> Running our services on computers is exactly what we as an industry did in the old-timey days.
There are times I long for the simpler approach from the "old-timey" days, but I digress...
> You've got availability concerns (power, network, hardware failure, etc), the networking to get a static IP in the place the laptop lives might be a bit tricky, but you can absolutely do this.
Availability is a good point :), static IP too. Thanks for the thoughts.
Networking would definitely be a concern and I 100% agree that programming for higher failure rates is a must.
I would certainly default to bare metal, likely a simple go binary running as a Job side-runner for a main prod server hosted in the cloud.
> Lastly, I would definitely consider using gitlab.com to host code and run a gitlab runner from home. The previous parts of what I wrote become moot. The runner just connects when it can and runs jobs.
Gitlab runner looks promising, though I am not sure if it's intended for work outside the git workflow.
You can do a POST with an api access token to gitlab.com to get a job to run with custom environmental variable settings. So, you don’t have to use git to start a job.
> You can just install Linux on them and put them in your wardrobe.
But they might not be silent enough to not disturb you when sleeping.
Agree 100%.
> Another con is that they aren’t rackmountable. So if you want to stack them you need to use some non-standard solution for that. And if a HDD/SSD fails, you need to order a replacement yourself, etc etc.
> It’s totally doable. But is it worth the hassle? That’s up to you.
Thanks for your unique take. HDD / SDD failing is definitely worth considering if it's worth the hassle. +1
I appreciate your response and don't necessarily disagree. Could you tell me a little more about why you wouldn't recommend it for a business, even if it is reliable?
> To get around the reliability problems, I have the on-prem laptop environment update cloudflare entries with keep-alive timestamp messages. .. I then have a Google script that monitors the keepalive and if too much time has passed, a “failover” is done and everything that was running on-prem is spun up in digital ocean. The failover script completes by updating DNS and pointing to DO
This is a clever solution. Thank you for your thoughts!