Hi, Im having problem with my fresh installation of opensuse (server minimal install). I have found some old bug related to this issue and swap, but I don’t have swap at all. I tried all the proposals, including shortening DefaultTimeoutStopSec=90s to 20s in systemd conf, commenting GRUB_TERMINAL=“console” in grub, but nothing works, I’m on nvme and boot is very fast, but every reboot or shutdown takes exactly 99 seconds, that’s so annoying :’(
Maybe someone experienced the same problem in here? How can I turn on detailed logs, or how can I look into existing logs, to see what is happening? journalctl shows only logs from actual boot, not what happened prior reboot. Thank you.
Any idea, why modern opensuse does not like old text console? :\ And I it must have been 90s timeout, sudo reboot to actual bios screen takes 9 seconds now.
I do not know why that change to the boot command line would affect boot time. I’m guessing that there was some other reason for the delay, but it happens to have cured itself.
There is a problem with the lvm2 services causing shutdown delays with no associated error logs. You can either disable all lvm2 services, or uninstall it (as long as you’re not using LVM, of course).
And reboots seem to be fast again. I hope it will not return after some longer uptime. When I stop them using systemctl stop, they’re stopped quickly, why are they terminating so long on reboot? I’m using btrfs on one of my partitions and I’m planning to use snapshots. Will I need these services for btrfs snapshots?
Btw. I’m not having them on my desktop machine with tumbleweed at all.
I don’t know what is causing the problem on some systems and not others. Not everyone here has experienced this problem, but it was happening on my laptop shortly after I upgraded to 15.1. I think it was also the cause of some random system freezes that would happen occasionally, and almost always when running a VM. Since disabling all of the lvm2 services, the issue has disappeared. Although, I also switched to the tumbleweed kernel so I’m not 100% certain if it was lvm2 or the older leap kernel or some combination of the two that was causing the problem. But, I have a functional system now so I’m not going to mess with it.
Since you have standard partitions, btrfs, ext4, etc., you can leave lvm2 disabled without affecting anything else on your system.
Thank you, you helped me a lot, I have fast reboots now. So I’ll leave lvm2 disabled for now, my OS15.1 is not running in VM, so I hope I won’t need lvm and maybe in 15.2 they will fix it. I was afraid to install tumbleweed on “production” system, but on my pc I’m happy with tumbleweed too.
lvm2 has nothing to do with VMs. LVM = Logical Volume Manager. In short has to do with disk management / use. And, if you don’t use LVM, it’s no problem to disable the services, even remove and lock the packages.
My impression, from reading these reports and comparing to what I see, is that this appears to be a problem for an LVM service that is not actually used.
Disabling lvme has some side effect though, if I do sudo update-grub (which actually runs that opensusish update boot-whatever which I never remeber)
sudo update-grub
Generating grub configuration file ...
Found linux image: /boot/vmlinuz-4.12.14-lp151.28.32-default
Found initrd image: /boot/initrd-4.12.14-lp151.28.32-default
Found linux image: /boot/vmlinuz-4.12.14-lp151.28.25-default
Found initrd image: /boot/initrd-4.12.14-lp151.28.25-default
ss WARNING: Failed to connect to lvmetad. Falling back to device scanning.
So I did sudo zypper rm -u lvm2 and update-grub is now without errors Thank you!
Sorry for a dumb question, but how do I surely know that I do not need it? I mean, say I’d create an encrypted lvm volume, then got a hit into the head, forgot that, and removed lvm2. Would opensuse allow me to do that (and then fail to boot), or is there some mechanism, which warns stupid users going to break their system, and warns thats it is in use? I remember ubuntu warning me I’m going to do somehing really stupid which will lead to breaking my system, when I tried to remove some important package.
I just tried removing lvm2 with Yast software management. It does not give any warnings.
So I clicked “Cancel” because I really didn’t want to make the system unbootable. It then gave me a warning that this would undo my selections for software changes. But never a warning that it would cripple my system.
If I really made that mistake, I think I could recover. I would have to boot from rescue media, and unlock the encrypted LVM there. And then go into a chroot session, where I would reinstall lvm2 and regenerate the “initrd”.
Thank you!