History

Versions 107

Version Author Diff Last updated
v107 Gabriel Filion
Wrong title for new section I forgot to change it after the copy/paste
v106 Gabriel Filion
new playbook for obsolete packages.
v105 Gabriel Filion
Note about what's updating the apt metrics and at which frequency
v104 Gabriel Filion
cross-reference yubikey touch policy from reboots documentation
v103 Gabriel Filion
Typo
v102 Gabriel Filion
Move up example for ganeti- + non-ganeti- reboots for full fleet This is the most likely scenario so it should appear first in the documentation. I've also modified the example command for running non-ganeti reboots so that it directly launches the reboots and it automatically excludes ganeti nodes.
v101 Gabriel Filion
reboot examples use old options names the options were renamed to include the unit type (e.g. minutes or seconds)
v100 Gabriel Filion
Example reboot commands fail If I just copy/paste the examples from this page, I get an error like this: No idea what 'perdulce.torproject.org' is! This is because the last sed command in subshells that fetch information is not doing what it should since output is separated by newlines instead of spaces. With `paste` we can make this work better.
v99 Antoine Beaupré
bullseye upgrades completed, congrats @lavamind and everyone! See tpo/tpa/team#41819
v98 Antoine Beaupré
more forgotten reboot job renames
v97 Antoine Beaupré
fix fab call syntax space-separated hosts don't work anymore
v96 Antoine Beaupré
upgrade global upgrade graph and stats
v95 Antoine Beaupré
deprecate ./reboot command We move towards the fleet task group, and more broadly away from those custom scripts. In particular, for reboots, we'd like to have a plain fleet.reboot command that would just do the right thing with all hosts.
v94 Antoine Beaupré
note the existence of canceling reboots
v93 Antoine Beaupré
clarify stats on the different releases
v92 Antoine Beaupré
regenerate all history graph with an area plot, fixing scale
v91 Antoine Beaupré
note flaw in X axis
v90 Gabriel Filion
Add note about cases caused by ganeti instances The easy and long methods. The long method reduces the amount of downtime for our VMs but is a bit longer to do, especially just to figure out which instances used a 'plain' disk template to handle them differently.
v89 Antoine Beaupré
more info about apt update issues (tpo/tpa/prometheus-alerts#22)
v88 Antoine Beaupré
fix typo