| ianclark001 | Hi all, today I found myself unable to boot (specifically to get past the full-disk encryption screen). I couldn't find a way to retrieve any meaningful information as to why this was. I'm able to boot using the recovery mode, and everything works fine when using 3.19.0-28-generic, but 3.19.0-30-generic just won't boot. | 08:39 |
|---|---|---|
| ianclark001 | Would be super-appreciative of any guidance as to how I should approach debugging / resolving | 08:39 |
| smb | Hm, if people would give a little time to respond... | 08:58 |
| caribou | apw: smb: do you guys have good ties with the Debian kernel people ? | 09:04 |
| caribou | I'm thinking of proposing my "small initrd" changes to Debian as well | 09:05 |
| caribou | I see that dannf is in the team, I'll probably ping him about this later | 09:05 |
| smb | caribou, We (ok, probably apw more) have talked to Ben a few times... | 09:06 |
| caribou | smb: on a side note, who "owns" the /boot directory ? | 09:07 |
| apw | caribou, i'd not call our ties strong no, they should be better than they are, i guess henrix talks to ben a fair bit on security topics | 09:07 |
| caribou | I mean who whould mostly be concerned about kdump-tools putting stuff there ? the kernel team ? | 09:07 |
| apw | between us and the bootloader people | 09:08 |
| smb | Maybe a shared interest between kernel and grub... | 09:08 |
| caribou | apw: I just don't want to ruffle feathers by adding stuff there | 09:08 |
| apw | did we decide it had to be in there, i thought they were loaded after we had filesystems | 09:08 |
| apw | else /var/lib/kdump or whatever is an option | 09:08 |
| caribou | apw: hmm, true; might be better there, let me test | 09:09 |
| caribou | apw: I need to test. kexec is mostly concerned there | 09:09 |
| apw | ack | 09:10 |
| caribou | apw: ok, thanks for the info | 09:10 |
| caribou | apw: smb: I will write a blog post about the proposed changes | 09:11 |
| apw | nice | 09:12 |
| caribou | apw: FYI - file system availability is not an issue here | 09:55 |
| caribou | the kexec call at boot time goes and reads the content of vmlinuz & initrd.img in memory for later use | 09:55 |
| caribou | hmm, well that's only true for the vmlinuz | 09:58 |
| caribou | nevermind, initrd.img goes in memory too | 10:02 |
| * caribou is making his way into kexec sources | 10:02 | |
| apw | caribou, right they are both loaded at "setup time" | 10:47 |
| apw | so its all about the kdump job, where it comes in the boot process, whether it is after wait for filesystems or not | 10:48 |
| caribou | apw: kdump-config load happens late at boot time | 10:49 |
| caribou | apw: it happens after network-online.target | 10:50 |
| apw | then i'd say it matters not where they are in a disk sense | 10:50 |
| caribou | apw: I had a doubt for a while on wether kexec was loading them only when triggered so I went back to the source to check | 10:50 |
| apw | caribou, no it definatly has to load them at init time, else it would rely on a lot of the kernel working to dump, which would defeat the purpose | 10:51 |
| caribou | apw: indeed | 10:51 |
| === shirgall_ is now known as shirgall | ||
| === MO is now known as Guest87877 | ||
| aisrael | I'm running 15.10 and noticed (after a failed reboot) that there's no packaged for the signed kernel for 3.16.0-41-generic (looks like the last signed is -30). Is there something I'm missing, or should I downgrade kernels (this is on a Mac Mini so the signed kernel is important) | 17:12 |
| apw | aisrael, do you have linux-signed-generic installed ? | 17:16 |
| aisrael | apw: Yes, sorry. I have 3.16.0-41-generic signed. I meant to say that there's no signed package for 3.19.0-30 | 17:17 |
| apw | aisrael, ok i am utterly confused, you are installinling linux-lts-vivid-generic ? | 17:18 |
| apw | aisrael, to swithc up from lts-utopic to lts-vivid ? | 17:19 |
| apw | or did you upgrade utopic to vivid ? | 17:19 |
| aisrael | apw: I upgraded from trusty -> utopic -> vivid | 17:19 |
| aisrael | I just found the signed package -- it wasn't installed when I upgraded last, so I might be wasting your time. :( | 17:20 |
| apw | aisrael, ok and do you have linux-signed-generic package installed | 17:20 |
| apw | as it is that that makes sure the signed kernels get installed as we release them | 17:20 |
| aisrael | apw: Aha. I did not have the meta package for the signed kernel installed | 17:20 |
| apw | and that would account for the lack of a signed one then | 17:21 |
| aisrael | I'll reboot that machine and verify that worked (and I'll add notes to the Ubuntu on Mac docs to note that requirement) | 17:21 |
| aisrael | apw: That fixed it, thank you! | 17:23 |
| lfaraone | jsalisbury: can you also upload a sig on the SHA256SUMs for your latest build in http://kernel.ubuntu.com/~jsalisbury/lp1500751/ ? :) | 18:14 |
| * lfaraone is paranoid, sorry. | 18:14 | |
| jsalisbury | lfaraone, sure | 18:30 |
| jsalisbury | lfaraone, done. In a file named: signature-lp1500751-commit-90dcba7ee | 18:33 |
| ogra_ | grmpf ... | 21:06 |
| ogra_ | does anyone know a way to make devtmpfs re-populate /dev ? i ran a broken script that wiped my dev dir and dont want to reboot | 21:07 |
| apw | ogra_, maybe you can retrigger udev to repopulate it | 22:15 |
| ogra_ | apw, i tried that ... it doesnt recreate the initial nodes | 23:12 |
| ogra_ | in the end i bit the bullet and rebooted | 23:12 |
| ogra_ | (i had a semi populated /dev ... but all block devices were missing etc) | 23:12 |
| === mamarley_ is now known as mamarley | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!