Bug #17458

Upgrade to Buster 10.3

Added by intrigeri 2020-02-04 07:06:32 . Updated 2020-02-10 10:11:42 .

Status:
Resolved
Priority:
High
Assignee:
Category:
Target version:
Tails_4.3
Start date:
Due date:
% Done:

100%

Feature Branch:
bugfix/17458-buster-10.3+force-all-tests
Type of work:
Code
Blueprint:

Starter:
Affected tool:
Deliverable for:

Description

It is scheduled for Saturday, February 8th, 2020.

Subtasks

Related issues

Blocks Tails - Feature #16209: Core work: Foundations Team Confirmed
Blocked by Tails - Feature #17443: Upgrade Linux to 5.4.8+ Resolved

History

#1 Updated by intrigeri 2020-02-04 07:06:52

#2 Updated by intrigeri 2020-02-04 10:25:27

  • Status changed from Confirmed to In Progress
  • Feature Branch set to bugfix/17458-buster-10.3+force-all-tests

As usual, I’ve enabled buster-proposed-updates on the topic branch so we get CI results and can inspect the .packages diff. If we’re happy and merge this, after Buster 10.3 is actually out, we’ll need to bump the debian APT snapshot and revert commit:807e04bebbc78d60c2dce80eb6831eb6cce3fff3.

This is blocked by Feature #17443 because if we only bumped the APT snapshot, the branch would FTBFS due to Linux 5.3 not being in sid anymore.

#3 Updated by intrigeri 2020-02-04 10:26:12

#4 Updated by intrigeri 2020-02-04 11:14:04

I’ve compared the resulting .packages with the one from the Feature #17443 branch and everything makes sense to me.
Let’s see how the test suite behaves!

#5 Updated by intrigeri 2020-02-05 07:06:32

  • Status changed from In Progress to Needs Validation
  • Assignee deleted (intrigeri)

Full test suite passed.

After merging this branch, please reassign to me, so that after Buster 10.3 is actually out, I bump the debian APT snapshot and revert commit:5e4b0b5806f5958ad25f4440dfe925f8f7a84180.

#6 Updated by anonym 2020-02-05 15:18:42

  • Assignee set to anonym

#7 Updated by anonym 2020-02-06 16:44:17

  • Status changed from Needs Validation to In Progress
  • Assignee changed from anonym to intrigeri
  • % Done changed from 0 to 90

intrigeri wrote:
> Full test suite passed.

Same for me, locally.

I built/reviewed/tested this at the same time as Feature #17443 (linux + APT snapshot bump) and Feature #17059 (tor bump), but after taking those into account the .packages diff looks right and not very scary. Merged!

> After merging this branch, please reassign to me, so that after Buster 10.3 is actually out, I bump the debian APT snapshot and revert commit:5e4b0b5806f5958ad25f4440dfe925f8f7a84180.

Done!

#8 Updated by intrigeri 2020-02-06 18:02:28

  • Priority changed from Normal to High

#9 Updated by intrigeri 2020-02-09 12:03:32

If the updated branch builds images that look good and the plymouth/cryptsetup regression @CyrilBrulebois mentioned turns out to be a false alert for our use case, I’ll need to do this:

ssh reprepro-time-based-snapshots@incoming.deb.tails.boum.org \
       tails-bump-apt-snapshot-valid-until debian 2020020902 120

#10 Updated by intrigeri 2020-02-09 15:52:59

  • Status changed from In Progress to Needs Validation
  • Assignee changed from intrigeri to anonym

Hi @anonym,

for 4.3 we really need to merge this or to revert the already-merged version of this branch. So:

intrigeri wrote:
> If the updated branch builds images that look good

The .packages list for the topic branch is identical to the one from a current stable branch build. With this in mind, I would normally just merge this, but this makes me wish a 2nd pair of eyes:

> and the plymouth/cryptsetup regression @CyrilBrulebois mentioned turns out to be a false alert for our use case,

I’ve got no news about this so far. Given our usage of plymouth (merely to display a splash screen between syslinux & GDM, plus displaying a nice error if X.Org fails to start), I doubt this is going to affect us in practice.

> I’ll need to do this:

Done, to ensure we don’t forget about it.

#11 Updated by anonym 2020-02-09 16:05:11

intrigeri wrote:
> […] I would normally just merge this, but this makes me wish a 2nd pair of eyes:
>
> > and the plymouth/cryptsetup regression @CyrilBrulebois mentioned turns out to be a false alert for our use case,
>
> I’ve got no news about this so far. Given our usage of plymouth (merely to display a splash screen between syslinux & GDM, plus displaying a nice error if X.Org fails to start), I doubt this is going to affect us in practice.

A full test suite run is on its way, and I’ll test booting on baremetal. Not sure what more I could do (?). I both of these looks good I’ll merge this early tomorrow.

#12 Updated by anonym 2020-02-10 10:11:42

  • Status changed from Needs Validation to Resolved
  • Assignee deleted (anonym)
  • % Done changed from 90 to 100

LGTM! Merged!