Package: libparted2-udeb Version: 3.2-1 Severity: serious Justification: what subject says
[X-D-Cc: -boot@, please keep it in the loop.] Hi, thanks for looking into #757417 and performing a number of uploads to sid. I've performed a few tests using those, and I can confirm that at least: - Full disc w/o LVM from a blank disc = OK - Full disc with auto LVM from the resulting disc = OK - Full disc with auto LVM from a blank disc = OK But starting w/ or w/o a blank disc I'm getting: - Full disc with auto encrypted LVM = FAIL Indeed I'm getting an error message: no partition has been selected for the root filesystem. I have no strong background on the whole partman-* ecosystem, but it looks to me like the partman-auto-crypto package might have needed an update similar to the one you just performed for the other packages. Since performing some testing is something I'm slightly more competent with, I've checked that the following images work OK for full disc auto encrypted LVM: - Jessie Alpha 1 netinst. - Same image as above (meaning udebs from sid, basically), but with parted udebs reverted to their 3.1-4 versions. Since you mentioned in a reply that the partman log alone should be sufficient I'm attaching the logs for: - sid only = FAIL (partman.sid-fail.txt.xz) - sid with reverted parted = OK (partman.sid+revert-ok.txt.xz) Looking at the first diff hunk, I might not be totally mistaken about a possibly missing free vs. unknown patch: -parted_server: OUT: 1 0-1780482047 1780482048 primary unknown /dev/mapper/sda5_crypt +parted_server: OUT: -1 0-1780482047 1780482048 primary free /dev/mapper/sda5_crypt [ I should be able to perform some testing as soon as you have a prospective patch or package, by the way. Hopefully that's going to be the last round trip before we get everyone into testing and cut a release. ] Mraw, KiBi.
partman.sid-fail.txt.xz
Description: application/xz
partman.sid+revert-ok.txt.xz
Description: application/xz