You are not logged in.
I just updated systemd from version 208-10 to 212-1. After a reboot, systemd waits for the availability of the mount unit files from the systemd-fstab-generator. After 90 seconds of waiting, it gives up and drops into the recovery shell. In the recovery shell, you can do
systemctl daemon-reload
systemctl default
For some reason, disk I/O becomes extremely slow in the recovery shell. I don't know if systemd enables some mount flags (such as sync) to make it so slow. A slightly more permanent workaround is to change your /etc/fstab so that you don't have any of the mount points automounted (except for / of course). You can do this with the noauto flag.
Is anyone else experiencing this issue?
Last edited by imran1008 (2014-04-07 04:05:14)
Offline
I am on systemd 212-2. Just rebooted fine. Do you have a nonstandard /etc/fstab?
gpg --keyserver hkps.pool.sks-keyservers.net --recv-key 0x77660e16dba99933
0B85 1AC2 4980 B654 0DA3 D1E1 7766 0E16 DBA9 9933
Offline
I don't think so. This happened to me on two of my machines. Here is one of my /etc/fstab:
tmpfs /tmp tmpfs nodev,nosuid 0 0
UUID=471f74d2-5a34-4e1e-b8c3-cfb0335afda5 / btrfs space_cache,inode_cache,compress=lzo 0 1
UUID=734a3085-a3c3-4775-9c92-91fbbe776823 swap swap defaults,pri=1 0 0
UUID=6bd77ebc-dc85-4a0e-bc3e-4aa6a5160178 /boot ext2 defaults 0 1
UUID=fecfa3d1-7889-4c0f-92a9-870457c19f39 swap swap defaults,pri=1 0 0
UUID=4340edf1-4f90-44a7-83b4-f8999180422f /mnt/tmp btrfs space_cache,inode_cache,compress=lzo 0 0
Sorry about alignment. I tried each one of them individually and got the same problem. Only after commenting out everything except for root (/) made the system boot without problems.
Offline
I also tried it with /dev/sda# notation (not just with UUID)
Offline
Found the issue! It was the kernel. I had CONFIG_AUDIT enabled in my kernel and the README file of systemd mentions that it doesn't work well with the option enabled. Disabling it fixed the issue. I also change a few other options, such as enabling CONFIG_FHANDLE, CONFIG_IPV6 (from module to built-in), CONFIG_AUTOFS4_FS (from module to built-in), CONFIG_SCHEDSTATS and disabling CONFIG_FW_LOADER_USER_HELPER. I highly suspect it was the disabling of CONFIG_AUDIT that fixed it.
Offline