Skip to content

Conversation

@dmnks
Copy link
Contributor

@dmnks dmnks commented May 9, 2024

Instead of storing the epoch per-repo, use the machine-id(5) file's timestamp as the source of truth. Please see the commit for details (GH didn't auto-fill the description here when creating the PR for some reason).

Note: Only intended for F41+ (should be reverted in F40 and lower).

Related test PR: rpm-software-management/ci-dnf-stack#1501

Fixes: #1611

dmnks added 2 commits May 9, 2024 10:39
The buckets aren't really an array that's indexed in the code, they're
just sequential numbers for the URL flag.  Also clarify why we're using
"this window" instead of "the current position of the sliding window" in
the comments.
Actually use the system's installation time (if known) as the reference
point, instead of the first-ever countme event recorded for the given
repo.

This is what the dnf.conf(5) man page always said about the countme
option, the code just never lived up to that.

This makes bucket calculation more accurate:

1. System upgrades will no longer reset the bucket to 1 (this used to be
   the case due to a new persistdir being created whenever $releasever
   changed).

2. Systems that only reach out to the repos after an initial time period
   after being installed will no longer appear younger than they really
   are.

3. Prebuilt OS images that happen to include countme cookies created at
   build time will no longer cause all the instances spawned from those
   images (physical machines, VMs or containers) to appear older than
   they really are.

Use the machine-id(5) file's mtime to infer the installation time.  This
file is semantically tied to the system's lifetime since it's typically
populated at installation time or during the first boot by an installer
tool or init system, respectively, and remains unchanged.

The fact that it's a well-defined file with clear semantics ensures that
OS images won't accidentally include a prepopulated version of this file
with a timestamp corresponding to the image build, unlike our own cookie
files (see point 3 above).

In some cases, such as in OCI containers without an init system running,
the machine-id file may be missing or empty, even though the system is
still used long-term.  To cover those, keep the original, relative epoch
as a fallback method.  System upgrades aren't really a thing for such
systems so the above point 1 doesn't apply here.

Some containers, such as those created by toolbox(1), may also choose to
bind-mount the host's machine-id file, thus falling into the same bucket
as their host.  Conveniently, that's what we want, since the purpose of
such containers is to blend with the host as much as possible.

Fixes: rpm-software-management#1611
@dmnks
Copy link
Contributor Author

dmnks commented May 9, 2024

Please note the countme test now fails in the CI since it's not compatible with the new code, that's fixed by rpm-software-management/ci-dnf-stack#1501.

Copy link
Member

@jan-kolarik jan-kolarik left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM, thanks!

@jan-kolarik
Copy link
Member

/packit test --identifier dnf-tests

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

Status: Done

Development

Successfully merging this pull request may close these issues.

Countme should report system age, not repository age

2 participants