Page MenuHomePhabricator

"discourse" Cloud VPS project jessie deprecation
Closed, ResolvedPublic

Description

The end of life of Debian Jessie is approaching in 2020 and we need to move to Debian Buster (or Stretch) before that date.

All instances in the discourse project need to upgrade as soon as possible. Instances not upgraded by 2019-12-31 may be subject to deletion unless prior arrangements for an extended deadline has been approved by the Cloud VPS administration team.

Remaining Debian Jessie instances (live report):

Listed administrators are:

See also:

More info on current project instances is available via openstack browser

Details

Due Date
Dec 31 2019, 11:59 PM

Event Timeline

StrikerBot created this task.

Ping @EBernhardson, @Qgil, @Samwilson, @Tgr, @Austin

The Cloud Services team would like to have Debian Jessie systems replaced with Debian Buster (or Stretch if necessary) before 2019-12-31. Please do respond of this task with comments if you know you will not be able to meet that target date or have additional questions about what to do or how to do it. Ideally you will create new instances in your Cloud VPS project, test them, and then migrate any final state data to the new instances before deleting the old Jessie instances. If you need more quota space in your project to create new instances in parallel with your existing instances please create a quota request task describing the increase you need and referencing this deprecation task.

Sorry for not answering before. This sounds like a good project for the last weeks of December, when "internal communications" slow down and there is more time and peace to do things.

Is there documentation available about the recommended way to proceed? Is this about

  1. Make a backup of all your data and take it out from the server.
  2. apt-get dist-upgrade

or is there more?

Doing apt-get dist-upgrade is technically possible, but we don't encourage it becuase we would lost track of what the base operating system is. This is something that could be improved in our side but that won't happen in the short term I'm afraid.

Ideally you will create new Debian Buster instances in your Cloud VPS project, test them, and then migrate any final state data to the new instances before deleting the old Jessie instances.

I would discourage enabling people to just dist upgrade more easily and
with support due to the implication that the instance is hard to replace
from scratch. People should be prepared to replace these. Especially
considering we can lose them irrecoverably.

Killing the old box and pulling up a new one is probably less complex and takes less time than a dist upgrade, anyway. Also less disruptive as the original server can be kept running until the new one has been sufficiently tested.

Ah, wait, now I realize that this upgrade only affects discourse1002.discourse.eqiad.wmflabs aka https://discourse.wmflabs.org/

Anyway I wanted to get rid of that Discourse instance. Its closure was announced months ago. I wanted to create a replica of Discuss Space in service of T235235: Document data backup and disaster recovery plan for https://space.wmflabs.org and https://discuss-space.wmflabs.org.

In this case I don't think we need the extra quota.

Qgil edited projects, added Discourse; removed Space (Oct-Dec-2019).

The instance discourse1002.discourse.eqiad.wmflabs has been deleted.

I have a backup stored in my computer... just for the record.