On Wed, Sep 07, 2016 at 10:36:29 -0300, Martín Marqués <martin.marques(a)gmail.com&gt; wrote: Those sound like systemd messages. You probably need to find out which dependency failed and that should be a big clueto which update broke things.

2016-09-07 11:01 GMT-03:00 Bruno Wolff III <bruno(a)wolff.to&gt;: Here's an oversite on my side: Warning: crypto LUKS UUID 9fe296b1-8aa9-4cd6-9868-b37e2720407b But looking at the /dev/mapper where the device is I see: /dev/mapper/luks-263bd743-5d1b-485b-9884-c3eda1a4d880 Looking at the /run/initramfs/rdsosreport.txt it's actually finding the device in the which I see in /dev/mapper, which is the correct UUID I have in /etc/fstab. So what is the UUID 9fe296b1-8aa9-4cd6-9868-b37e2720407b I see in the output from dracut? -- Martín Marqués select 'martin.marques' || '@' || 'gmail.com' DBA, Programador, Administrador

On Wed, Sep 7, 2016 at 1:00 PM, Chris Murphy <lists(a)colorremedies.com&gt; wrote: Do get more information from these failed boots, add this boot parameter: systemd.log_level=debug Post the resulting rdsosreport.txt somewhere, or you can sift through and find out a little more about why it's failing. There's something looking for a UUID that's not present, so we have to figure out what's asking for that UUID. Why this coincides with updates, I have no idea yet so far. There's no reason why a UUID would change with an update... well except.... Get a load of this weirdness. So I have a Fedora Server, with encrypted swap. It's using /dev/urandom at boot to get a new key everytime, so I don't have to plug in a passphrase at boot time. In the /etc/crypttab I'm using /dev/disk/by-id/wwn-<someserialnumberorportionthereof>-part5 to identify the exact partition. Works great. Then I did an update one day and boot starts failing waiting on this device that doesn't exist anymore. Guess what? The update changed how it was creating the /dev/disk/by-id/wwn- Before the update, it was bass ackwards, something like last 4+middle 4+middle 4 part 2+first 1 of the unique ID+the NAA+3 zeros (?). Whatever it was it was produced from the WWN but didn't match it exactly. After the update, it matches it exactly. So maybe you're using WWN somewhere? *shrug* -- Chris Murphy