# Roll call: who's there and emergencies
anarcat, gaba, groente, lavamind, lelutin, zen.
There's significant noise in monitoring, but nothing that makes it worth canceling this meeting.
# Dashboard review
## Normal per-user check-in
Tried to make this section quick, but there were some discussions to be had:
* Development · Boards · The Tor Project · GitLab
* Development · Boards · The Tor Project · GitLab
* Development · Boards · The Tor Project · GitLab
* Development · Boards · The Tor Project · GitLab
* Development · Boards · The Tor Project · GitLab
## General dashboards
Skipped this section.
* Development · Boards · The Tor Project / TPA / TPA team · GitLab
* Development · Boards · Web · GitLab
* Development · Boards · TPA · GitLab
# Tails merge discussion
Let's review the work Zen did. Our rough plan was:
- confirm already identified consensus
- try to establish consensus on remaining items, or at least detail
controversies and blockers
- establish what should be done in 2025, 2026, < 2030, > 2030
We followed the [TPA-RFC-73 Draft](tpa rfc 73 tails infra merge roadmap · Wiki · The Tor Project / TPA / TPA team · GitLab) as it was at the time the meeting started.
We figured that today, we would agree on strategy (e.g. puppet merge), on the colors (e.g. which services are retired), and postpone the "what happens when" discussion. We also identified that most services above "low complexity" will require their own discussions (e.g. "how do we manage the Puppet control repo", "how do we merge weblate") that will happen later.
## Per service notes
- Alternative to puppet merge: migrate services to TPA before moving Puppet, but not a good idea because some services can't be easily migrated.
- registrars and colo could just depend on password store and not be otherwise changed.
- website depends on powerdns
- agreement of merging puppet codebases first
- eyaml: merge for now, until people get familiar with both trocla and eyaml, but we probably should have a single system for this
- virtualization: proposal: treat the old stuff as legacy and don't create new VMs there or make new hosts like those, if we need to replace hardware we create a ganeti box
- weblate:
- option 1: move the tor weblate to the self-hosted instance, need approval from emmapeel, check what reasons there were for not self-hosting
- option 2: move tails translation to tor's weblate and rethink the translation workflow of tails
We didn't have time to establish a 2025 plan, and postponed the rest of the discussions here.
# 2025 roadmap brainstorm
Throw ideas in the air and see what sticks about what we're going to do
in 2025. Following, of course, priorities established in the Tails
roadmap.
Postponed.
## What we promised OTF
For Tails:
- B.2: Keep infrastructure up-to-date and secure
As in Year 1, this will involve the day-to-day work needed to keep the
infrastructure we use to develop and distribute Tails up-to-date. This
includes our public website, our development servers for automatic builds
and tests, the translation platform used by volunteers to translate Tails,
the repositories used for our custom Debian packages and reproducible
builds, etc. Progressively over Year 2 of this contract with OTF, as Tails
integrates within the Tor Project, our sysadmins will also start maintaining
non-Tails-specific infrastructure and integrate internal services offered by
Tails within Tor’s sysadmin workflow
Nextcloud
For TPA:
- I didn't find anything specific for TPA.
# Metrics of the month
* hosts in Puppet: 90, LDAP: 90, Prometheus exporters: 504
* number of Apache servers monitored: 34, hits per second: 612
* number of self-hosted nameservers: 6, mail servers: 11
* pending upgrades: 0, reboots: 77
* average load: 1.03, memory available: 3.50 TiB/4.96 TiB, running processes: 321
* disk free/total: 65.69 TiB/139.85 TiB
* bytes sent: 423.32 MB/s, received: 270.22 MB/s
* planned bookworm upgrades completion date: 2024-10-02
* [GitLab tickets]: 256 tickets including...
* open: 2
* icebox: 162
* future: 39
* needs information: 4
* backlog: 27
* next: 11
* doing: 5
* needs review: 7
* (closed: 3760)
[Gitlab tickets]: Development · Boards · The Tor Project / TPA / TPA team · GitLab
Upgrade prediction graph lives at bookworm · Wiki · The Tor Project / TPA / TPA team · GitLab
Note that we have only a single "buster" machine left to upgrade after the Mailman 3 upgrade, and also hope to complete the bookworm upgrades by the end of the year. The above "in 3 weeks" date is unrealistic and will be missed.
The "all time" graph was also rebuilt with histograms, making it a little more readable, with the caveat that the X axis is not to scale:
···
--
Antoine Beaupré
torproject.org system administration