Loading policy/tpa-rfc-73-tails-infra-merge-roadmap.md +48 −0 Original line number Diff line number Diff line Loading @@ -12,8 +12,47 @@ discussion: Summary: Tails infra merge roadmap. # Background In 2023, Tor and Tails started discussing the possibility of a merge and, in that case, how the future of the two infrastructures would look like. The organizational merge happened in July 2024 with a [rough idea](/tpo/tpa/team/-/wikis/roadmap/tails-merge) of the several components that would have to be taken care of and the clarity that merging infrastructures would be a several-years plan. This document intends to build on the work previously done and describe dependencies, milestones and a detailed timeline containing all services to serve as a basis for future work. # Proposal ## Goals ### Must have - A list of all services with: - a description of the service and who are the stakehoders - the action to take - the complexity - a list of dependencies or blocks - a time estimation - A plan to merge the Puppet codebases and servers - A list of milestones with time estimates and and indication of ordering ### Non-Goals - We don't aim to say exactly who will work on what and when ## Scope This proposal is about: - all services that the Tails Sysadmins currently maintain: each of these will either be kept, retired, merged with or migrated to existing TPA services (see the [terminology](#actions) below), depending on several factors such as convenience, functionality, security, etc. - some services maintained by TPA that may act as a source or target of a merge, or migration. ## Terminology ### Actions Loading Loading @@ -960,6 +999,15 @@ flowchart TD class Website merge ``` # Alternatives considered - **Converge both codebases before merging repositories and Puppet Servers:** This approach would have the following disadvantages: - keeping two different Puppet codebase repositories in sync is more prone to errors and regressions, - no possibility of using exported resources would make some migrations more difficult (eg. Backups, Monitoring, TLS, etc) # References See the [TPA/Tails sysadmins overview][] document that was used to inform the Loading Loading
policy/tpa-rfc-73-tails-infra-merge-roadmap.md +48 −0 Original line number Diff line number Diff line Loading @@ -12,8 +12,47 @@ discussion: Summary: Tails infra merge roadmap. # Background In 2023, Tor and Tails started discussing the possibility of a merge and, in that case, how the future of the two infrastructures would look like. The organizational merge happened in July 2024 with a [rough idea](/tpo/tpa/team/-/wikis/roadmap/tails-merge) of the several components that would have to be taken care of and the clarity that merging infrastructures would be a several-years plan. This document intends to build on the work previously done and describe dependencies, milestones and a detailed timeline containing all services to serve as a basis for future work. # Proposal ## Goals ### Must have - A list of all services with: - a description of the service and who are the stakehoders - the action to take - the complexity - a list of dependencies or blocks - a time estimation - A plan to merge the Puppet codebases and servers - A list of milestones with time estimates and and indication of ordering ### Non-Goals - We don't aim to say exactly who will work on what and when ## Scope This proposal is about: - all services that the Tails Sysadmins currently maintain: each of these will either be kept, retired, merged with or migrated to existing TPA services (see the [terminology](#actions) below), depending on several factors such as convenience, functionality, security, etc. - some services maintained by TPA that may act as a source or target of a merge, or migration. ## Terminology ### Actions Loading Loading @@ -960,6 +999,15 @@ flowchart TD class Website merge ``` # Alternatives considered - **Converge both codebases before merging repositories and Puppet Servers:** This approach would have the following disadvantages: - keeping two different Puppet codebase repositories in sync is more prone to errors and regressions, - no possibility of using exported resources would make some migrations more difficult (eg. Backups, Monitoring, TLS, etc) # References See the [TPA/Tails sysadmins overview][] document that was used to inform the Loading