Upgrade to Ubuntu 26.04 LTS from 22.04

[jpmckinney]

22.04 is EOL Jun 2027. https://ubuntu.com/about/release-cycle

Upgrade servers: ./run.py '*' cmd.run 'lsb_release -a'

• kingfisher-main (Hetzner Dedicated): https://ocdsdeploy.readthedocs.io/en/latest/deploy/servers/data-support.html#create-a-data-support-main-server
• dream-bi (Hetzner Cloud): First time migrating. Double-check files to copy with @ndrhzn, and korsikov on Slack, and then document these on a new docs page.
• prometheus (Linode): https://ocdsdeploy.readthedocs.io/en/latest/deploy/servers/prometheus.html#migrate-from-an-old-server
• docs (Linode): https://ocdsdeploy.readthedocs.io/en/latest/deploy/servers/docs.html#migrate-from-an-old-server
• cms (Linode): First time migrating entire server, though we recently migrated the corporate site. We should create a page like those above, building off the steps in https://docs.google.com/document/d/1HyNXRgzLu-x6pHDQ4hkp5vEbvUm6S29lyNbMTXeO4yY/edit?tab=t.0#heading=h.hv00oifa1gte

@jpmckinney: Change the Python versions in relevant GitHub Actions workflows and other files. 26.04 ships with 3.13. Some apps create a venv using the default Python version:

• collect-generic
• kingfisher-collect and bi.open-contracting.org/powerbi/Dockerfile_python
• kingfisher-summarize

@yolile Last time we migrated kingfisher-main, we didn't keep any of the Process, Summarize or Pelican data, or any files in local-load directories. We only kept the incremental crawl data. Do we want to do anything differently this time? (Slack)

[RobHooper]

The current estimated release of 26.04 is 23 Apr 2026.
We should wait for the production ready release (".1") before migrating server. This is usually released in August / summer time.
We can start updating and reviewing the core SaltStack for any feature changes on the new release.

[yolile]

@yolile Last time we migrated kingfisher-main, we didn't keep any of the Process, Summarize or Pelican data, or any files in local-load directories. We only kept the incremental crawl data. Do we want to do anything differently this time?

Yes, I think we should preserve Kingfisher Process data. We often need that one to compare if previous publications have improved (we are at the beginning of the year, so this is very relevant). We can re-run the others (Summarize, Pelican) as needed

[jpmckinney]

Sounds good, @yolile. We can check with the rest of the data team on Slack when we're closer to the migration date.