Nokogiri: Rebuild ci.nokogiri.org

[flavorjones]

Current configuration is expensive, and the latest concourse-deployment and bbl have options that are cheaper. Also, not sure I can admin the cluster anymore without lost creds.

[flavorjones]

See https://github.com/cloudfoundry/bosh-bootloader/blob/master/docs/concourse.md

And https://github.com/concourse/concourse-bosh-deployment

[flavorjones]

Local config:

$ bosh -v
version 5.2.2-d737b7d3-2018-08-28T16:56:35Z

$ bbl -v
bbl 6.9.16 (linux/amd64)

[flavorjones]

I've lost access to the current environment and it looks like I may have lost a few months of data. :-(

[flavorjones]

I've enlisted help from @osis who may be able to recover build history.

[osis]

Howdy!

Was able to grab the snapshot, spin up a new VM, attach mount the snapshot as disk, install Postgres, and pgdump to victory! 🎉

Hoping to see the site up again soon!

[flavorjones]

All data recovered thanks to @osis. You rock! <3 <3 <3

[flavorjones]

Basic pipelines up and running again at https://ci.nokogiri.org/

Still working on getting workers stable.

[flavorjones]

I think everything's mostly working, except for:

• [ ] windows workers
• [x] scale up all workers
• [x] container placement strategy
• [x] x-frame options

And I haven't tested this yet:

• [x] PR triggers / status
• [x] IRC notifications

Should get to it all in the next day or so.

[flavorjones]

https://github.com/concourse/concourse-bosh-deployment/pull/99

[flavorjones]

https://github.com/concourse/concourse-bosh-deployment/pull/100

[flavorjones]

Closing this. Windows workers are blocked on https://github.com/sparklemotion/nokogiri/issues/1729