Since systemd has made it's way onto my computer, booting has become quite painful.... Here's what happens:
I get dropped to a recovery shell because systemd was unable to mount all my data partitions. I have 2 LVM VG's; one on my SSD where the 'system' lives (/, /var and /usr), and a second on a software RAID-1 with various data mounts. The 'SSD' VG gets activated and all mounts in there are mounted OK. In the recovery console, I have to 'pvscan', 'vgchange -ay' and 'mount -a' to get everything mounted. I still get errors about not being able to mount networked filesystems (it seems networking hasn't been brought up yet) but I'm able to continue booting.
So it seems that my second VG isn't being activated properly; note that I don't have to enable the software RAID, it is already running, I just have to activate the VG.
Once I have done all that, I can run 'systemctl default' to get to my login screen, but it takes around 3 to 4 minutes to do that.
I have no idea where to start looking now that we have systemd, so can anyone give me some suggestions for debugging why everything isn't mounting, and why it takes so damn long?
Last edited by fukawi2 (2013-04-01 00:14:54)
This is likely to do with the move to using lvmetad in the lvm2 package. This inability to find the rootfs on boot is something I have been facing on and off for a little bit now. Though I don't think it has happened at all with the last couple Arch kernels. But this doesn't mean much, as it wasn't happening with graysky's ck kernel for a couple kernels, and is now happening regularly again with 3.8.5.
So I don't really know what to tell you on that. I read somewhere that someone who compiled the lvm2 from git said that this seems to have been fixed there. But I am somewhat skeptical since I have experienced such ups and downs with my system.
I can tell you though that I really do not think this has anything to do with systemd, as lvm2 had been working flawlessly on my system for a while since I started using it and systemd together. This is certainly a recent occurance.
Edit: Check this thread out: https://bbs.archlinux.org/viewtopic.php?id=158012 (I know the title indicates LVM and Luks, but it is really a thread about LVM)
Last edited by WonderWoofy (2013-03-30 06:15:45)
It's certainly only been a problem since I (finally) upgraded to systemd.
I never got this news article in my email which is weird, but I've followed it and it was all already done anyway.
Open to other ideas, getting very frustrating now that my system has started segfaulting and I'm having to reboot regularly.
Just a suggestion. Maybe you should remove the [systemd] from your title. Systemd is pretty much assumed around here these days, and i mistook it as solved several times.
OK, so there seems to be a few issues here.... After booting a rescue CD, I was still getting kernel panics. Seems like the RAM I just upgraded is faulty since all seems good after changing back to my old RAM.
Secondly, after chroot'ing in a updating everything, including new kernel and rerunning mkinitcpio I can't reproduce the problem with only 1 of the 2 VG's being activated. I did however still get dumped to the recovery console, which after some digging around, it seems that "Failed at step EXEC spawning /bin/plymouth: No such file or directory" actually means I still have run level numbers in my syslinux config still (ie, 5 for graphical, 3 or CLI only)... Nothing like helpful error messages: https://bugs.archlinux.org/task/34345