| View previous topic :: View next topic |
| Author |
Message |
freifunk_connewitz
Apprentice
Joined: 08 Feb 2006
Posts: 231
|
 Posted: Thu Nov 02, 2023 2:01 pm Post subject: [workaround] luks-device not recognized via UUID anymore |
 |
|
Hi,
rebooting after yesterdays updates left me with a failing decryption by dmcrypt: the luks-encrypted usb device that normally gets unlocked and mounted during boot failed to do so. The device is configured in /etc/conf.d/dmcrypt like this:
.
It gets mounted automatically according to an entry in /etc/fstab pointing to the unlocked partition under /dev/mapper/xyz.
But now the boot process halts with a dmcrypt error message: "Device /dev/sdb1 already contains a superblock signature." and a warning that continuing would erase data. If I type YES, the password prompt shows up, after entering the password fsck on the unlocked device fails and mounting as well.
The normal behaviour until yesterday: request password>unlock>fsck>mount.
After some testing I found out the problem lies in the UUID determination of the partition. When I issue manually:
| Code: | | cryptsetup luksOpen UUID=XXXXX crypt-usb |
dmcrypt fails with "UUID=XXX ist not a valid Luks device".
But when I use the dev location it works without problems.
So the solution to my problem was to change the entry in /etc/conf.d/dmcrypt from
to
(although many manuals say one should avoid the automatically enumbered dev names, I know).
Just in case somebody else has the same problem.
Since the only change to my system was a world update, probably one of these packages is behind this regress:
| Quote: | genlop --list --date yesterday
* packages merged:
Wed Nov 1 11:40:48 2023 >>> sys-apps/hwdata-0.374
Wed Nov 1 11:41:26 2023 >>> media-libs/libraw-0.21.1-r1
Wed Nov 1 11:41:34 2023 >>> app-office/libreoffice-l10n-7.5.6.2
Wed Nov 1 11:42:33 2023 >>> sys-apps/systemd-utils-254.5-r1
Wed Nov 1 11:42:58 2023 >>> net-dns/libidn2-2.3.4-r1
Wed Nov 1 11:43:06 2023 >>> dev-python/typing-extensions-4.8.0
Wed Nov 1 11:43:14 2023 >>> dev-python/zipp-3.17.0
Wed Nov 1 11:43:40 2023 >>> sys-apps/file-5.45-r3
Wed Nov 1 11:44:17 2023 >>> dev-libs/libsodium-1.0.19-r1
Wed Nov 1 11:44:31 2023 >>> www-plugins/chrome-binary-plugins-118.0.5993.117
Wed Nov 1 11:44:40 2023 >>> dev-python/setuptools-scm-8.0.4
Wed Nov 1 11:44:49 2023 >>> dev-python/trove-classifiers-2023.9.19
Wed Nov 1 11:45:03 2023 >>> dev-python/cffi-1.16.0
Wed Nov 1 11:45:41 2023 >>> dev-python/cryptography-41.0.4
Wed Nov 1 11:45:50 2023 >>> dev-python/urllib3-2.0.6
Wed Nov 1 11:47:31 2023 >>> media-libs/harfbuzz-8.2.0
Wed Nov 1 11:47:47 2023 >>> media-fonts/urw-fonts-20230503-r1
Wed Nov 1 11:49:00 2023 >>> x11-base/xorg-server-21.1.9
Wed Nov 1 11:49:15 2023 >>> x11-drivers/xf86-input-libinput-1.4.0
Wed Nov 1 11:49:57 2023 >>> x11-drivers/xf86-video-intel-2.99.917_p20230201
Wed Nov 1 11:50:17 2023 >>> net-im/signal-desktop-bin-6.35.0
Wed Nov 1 11:50:30 2023 >>> www-client/opera-104.0.4944.36
Wed Nov 1 11:50:50 2023 >>> www-client/vivaldi-6.2.3105.58
Wed Nov 1 11:52:50 2023 >>> app-text/poppler-23.09.0
Wed Nov 1 11:55:31 2023 >>> dev-libs/protobuf-21.12
Wed Nov 1 11:55:59 2023 >>> sys-apps/iproute2-6.5.0
Wed Nov 1 11:56:32 2023 >>> app-crypt/keysmith-23.04.3
Wed Nov 1 11:57:43 2023 >>> app-crypt/gpgme-1.22.0
Wed Nov 1 11:58:04 2023 >>> sys-apps/smartmontools-7.4
Wed Nov 1 11:58:16 2023 >>> dev-java/openjdk-bin-17.0.8.1_p1
Wed Nov 1 11:58:26 2023 >>> dev-java/openjdk-bin-8.382_p05
Wed Nov 1 14:30:31 2023 >>> app-office/libreoffice-7.5.6.2
Wed Nov 1 14:34:45 2023 >>> media-video/ffmpeg-6.0-r9
Wed Nov 1 14:35:05 2023 >>> mail-client/thunderbird-bin-115.4.1
Wed Nov 1 14:35:38 2023 >>> net-im/zoom-5.16.5.303
Wed Nov 1 14:35:58 2023 >>> www-client/firefox-bin-119.0-r2
Wed Nov 1 14:46:19 2023 >>> media-libs/tg_owt-0_pre20230428
|
|
|
| Back to top |
|
 |
grknight
Retired Dev
Joined: 20 Feb 2015
Posts: 1662
|
| Posted: Thu Nov 02, 2023 2:50 pm Post subject: |
|
|
I doubt this script has changed at all in years.
Your manual test is faulty because, very early on, dmcrypt service executes blkid -l -t "${source}" -o device when given a UUID=, or PARTUUID=, etc. to translate to real device.
Then it does cryptsetup isLuks ${source_dev} to figure out what if it should open the device or create a plain crypt (e.g. for swap)
Who knows what is going on in the time. If there are other options involved or if it is just some race condition. |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Networking & Security
