Commit Graph

8 Commits

Author SHA1 Message Date
Itxaka
b0b2518e6a
Add label to luks partition and dont gate on label (#471)
* Add label to luks partition and dont gate on label

Instead of gating on labels, lets just add the label to the luks
partition, the same way we do to the underlying unlocked partition, so
they share the fs label. That way, the locking and unlocking refer to
the same label always

---------

Signed-off-by: Itxaka <itxaka@kairos.io>
2025-03-13 10:34:23 +01:00
Itxaka
74eb5903b3
bump yaml from v1 to v3 (#417)
* Fix job to run ginkgo from go.mod

Signed-off-by: Itxaka <itxaka@kairos.io>

* Bump yaml from v1 to v3

Signed-off-by: Itxaka <itxaka@kairos.io>

---------

Signed-off-by: Itxaka <itxaka@kairos.io>
2024-07-15 19:53:00 +02:00
Itxaka
166b8802ea Drop kairos lib and use sdk instead
Signed-off-by: Itxaka <itxaka.garcia@spectrocloud.com>
2023-06-14 15:03:31 +02:00
Itxaka
36f861fd0c Use reproducible UUIDs for partition identification
Signed-off-by: Itxaka <itxaka.garcia@spectrocloud.com>
2023-05-04 12:07:21 +02:00
Mauro Morales
0f3eee7851
Add golint
Signed-off-by: Mauro Morales <mauro.morales@spectrocloud.com>
2023-03-29 17:05:51 +02:00
Mauro Morales
f977f72e8d
Switch to using new collector pkg
Signed-off-by: Mauro Morales <mauro.morales@spectrocloud.com>
2023-03-29 17:27:55 +03:00
mudler
71506ff020 🌱 Remove challenger traces from here
The challenger has no business logic in here, and kcrypt should just
read what it needs. This goes in parallel with https://github.com/kairos-io/kcrypt-challenger/pull/10
Not blocking each other.

Signed-off-by: mudler <mudler@c3os.io>
2023-01-19 12:28:46 +01:00
Dimitris Karakasilis
018322ff00
Move from partition_info to config package
Now the kcrypt configuration will be just a block (`kcrypt`) in the regular
kairos configuration files.

We will use the config mechanism of looking up in multiple directories
becase the old code was only looking at `/oem` but at boot time
(initramfs) the configuration was present in `/sysroot/oem`.

The reason it "worked" was that we had a bug and a sealedvolume was
matching our request simply because we sent an empty UUID/Label/Name
and the sealedvolume had and empty UUID (empty matches empty, right?)

With that fixed on the challenger server side, it became obvious that we
never actually read the partition_info file because we were looking for
it at the wrong place.

Signed-off-by: Dimitris Karakasilis <dimitris@karakasilis.me>
2023-01-18 14:55:25 +02:00