SoylentNews is people
SoylentNews
NCommander (2)
michael@casadevall.pro
http://soylentnews.org/
I'm Michael, local administrator and madman on SoylentNews. I'm an adventurer, writer, and developer who lives out of his backpack travelling the world.
Journal of NCommander (2)
I intended to have a main page post following last Saturday's post about SN's averted shutdown, but after three attempts of trying to write it and it basically kept coming out as a copy and paste of the last posts I've written, I'm going to put that aside for the moment. Instead of simply repeating the same thing over and over, I'm just going to write a quick summary of what's been done
I've spun up a new development Linode account, and kolie put together the skeleton of an ansible playbook. I'm going to start hammering on it more tomorrow. The intent is to create a full staging environment for every aspect of the site. The playbook will also let us easily be able to do development locally as it will perfectly match what we're doing in production. That way things like configuration updates or even rehash code updates can be tested, deployed, and easily rolled back.
The first target is to rebuild the nginx terminator, as well as to use Traefik to help handle routing and termination. I re-installed the SSL terminator last November so its exact configuration is known and easy to recreate. Each component of the stack is going to get documented both via a playbook a wiki page on GitHub. After this, we'll start migrating services one by one until we can fully decommission the current VMs.
This also has the advantage that we will have something resembling change management, disaster recovery, and not deal with some strange part of the stack breaking when a configuration is updated which is always a good thing. It also means doing a large overhaul to rehash at least moves into the realm of "theorically possible", up from "impossible".
I'll try to keep people apprised of how things are progressing here every few days. I have no idea how long any of this will take, and even after the infrastructure rebuild is done, there's still a lot more that has to follow.
(Score: 5, Insightful) by Reziac on Wednesday May 31, @12:36PM
Thank you, it's nice to be kept informed. No need to write up anything formal, just "notes of the day" would be great.
And there is no Alkibiades to come back and save us from ourselves.
(Score: 5, Touché) by janrinok on Wednesday May 31, @01:30PM (2 children)
Great - but I cannot understand why it is being done.
(Score: 1) by redback on Wednesday May 31, @10:13PM (1 child)
because the current tech is a nightmare, and NC will need to help if they have any hope of rebuilding.
(Score: 2) by janrinok on Thursday June 01, @05:02AM
The current tech is running the site now. For the moment it works, as it has done for several years except for when NCommander crashed it - twice. But It needs changing. I agree.
NCommander is leaving. What the site looks like in 1 week, 1 month,1 year, 5 years will not affect him at all.
Kolie can rebuild the tech with his own team of trusted staff. If he decides that the existing team are not part of his plans then so be it; I will accept that decision - but let me get on with my life too. I am locked out of the current system so cannot do all of my job.
However, NCommander dictating how the future site will look or be structured is simply unnecessary. He stated that he hasn't got the time nor the interest to continue with the project. He should, IMO, stop creating new tasks and go then. I wish him no ill will - I have already said that we are indebted to him for what he has done. But let the team that will take the site into the future, whoever they are, get on with the job.
(Score: 1) by shrewdsheep on Wednesday May 31, @04:28PM
for making this last effort to make the site sustainable.