You are not logged in.
Hi,
I switched from ubuntu to Archlinux two weeks ago and about half of my boot attempts fail because of some problems with my lvm-volumes
My setup
I have my root partition on /dev/sda2 and an lvm partition /dev/sda3 containing logical volumes for /var /home and /boot
If booting succeeds these volumes appear as
/dev/mapper/VolGroup01-varVol
/dev/mapper/VolGroup01-homeVol
/dev/mapper/VolGroup01-bootVol
If my booting fails, My screen shows
startin version 218
/dev/sda2: clean, blabla files, blabla blocks
A start job is running for LVM PV scan on devices 8:3
This last message (together with 3 start jobs for dev-disk-by-x2duuid-SOMENUMBERS) appears for 90 seconds and then emergency mode is started
After logging in for maintenance /dev/mapper only contains "control" but not my logical volumes
module dm_mod is loaded
dmesg does not seem to show anything related to lvm or sda3
and /var/log contains only empty files (btmp, lastlog, wtmp)
Where can I find informations about this problem?
Any suggestions how to solve this problem?
Last edited by DaAlx (2015-04-06 12:45:17)
Offline
Hi,
I switched from ubuntu to Archlinux two weeks ago and about half of my boot attempts fail because of some problems with my lvm-volumes
My setup
I have my root partition on /dev/sda2 and an lvm partition /dev/sda3 containing logical volumes for /var /home and /boot
If booting succeeds these volumes appear as
/dev/mapper/VolGroup01-varVol
/dev/mapper/VolGroup01-homeVol
/dev/mapper/VolGroup01-bootVolIf my booting fails, My screen shows
startin version 218
/dev/sda2: clean, blabla files, blabla blocks
A start job is running for LVM PV scan on devices 8:3This last message (together with 3 start jobs for dev-disk-by-x2duuid-SOMENUMBERS) appears for 90 seconds and then emergency mode is started
After logging in for maintenance /dev/mapper only contains "control" but not my logical volumes
module dm_mod is loaded
dmesg does not seem to show anything related to lvm or sda3
and /var/log contains only empty files (btmp, lastlog, wtmp)Where can I find informations about this problem?
Any suggestions how to solve this problem?
Hello,
Disclaimer: Without any logs to see what is happening, this is just an educated guess because it sounds like what happened to me. Try getting the logs during the emergency shell process.
I've run into an issue myself with LVM. From what I understand, the pvscan processes get stuck waiting for a udev sync that no longer exists because udev is restarted while pvscan processes are running in the background. There hasn't been an official solution but I modified udev hook with this (can be found in /usr/lib/initcpio/hooks/udev):
while [ "$(pgrep -f pvscan)" ]; do
sleep 0.1
done
which forces udev to wait until the LVM processes finish before telling it to die. It is hacky and poor for a long term solution because we have lvm code in the udev hook is bad design but I haven't had that issue any more.
You can also try
In /etc/mkinitcpio.conf, you must change your HOOKS list:
HOOKS="base udev autodetect modconf block lvm2 filesystems keyboard fsck" is a hook list for script
HOOKS="base systemd autodetect modconf block sd-lvm2 filesystems keyboard fsck" is a hook list for systemd
but that stopped working for a few people on 3.19.
Here is the bug report to keep you up to date:
https://bugs.archlinux.org/task/41833#comment133971
Last edited by SirMyztiq (2015-04-06 00:36:46)
Offline
Thank you for this quick help. In fact I first tried out your second tip: Using the sytemd hook instead of the udev hook seems to work perfectly well for me (at least for the about 10 boots I did so far)
Offline