Description of problem: /usr and /var are on same partition as root partition, or '/' Booting soon after roll out of kernel-server-3.2.1-1.mga2-1-1.mga2 and of kernel-server-3.2.1-2.mga2-1-1.mga2, in each case with current Cauldron (and so latest systemd from that date and time), I have a non-running syslog: # systemctl status syslog.target syslog.target - Syslog Loaded: loaded (/lib/systemd/system/syslog.target; static) Active: inactive (dead) # systemctl start syslog.target # systemctl enable syslog.target Warning: unit files do not carry install information. No operation executed. # systemctl status syslog.target syslog.target - Syslog Loaded: loaded (/lib/systemd/system/syslog.target; static) Active: active since Sat, 21 Jan 2012 12:07:09 +0000; 19s ago
Blocks: (none) => 2120
Blocks: 2120 => (none)Summary: Since last 2 kernels (and boots) systemd deos not start syslog.target => Since last 2 kernels (and boots) systemd does not start syslog.target
Sorry for bad browswer/bugzilla interaction :((
Confirmed: seen on 2 Cauldron machines now.
do you have /etc/systemd/system/syslog.service ?
CC: (none) => dmorganec
Certainly; since months: stat /etc/systemd/system/syslog.service File: `/etc/systemd/system/syslog.service' -> `/lib/systemd/system/rsyslog.service' Size: 35 Blocks: 0 IO Block: 4096 symbolic link Device: 805h/2053d Inode: 411859 Links: 1 Access: (0777/lrwxrwxrwx) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2012-01-24 19:40:41.151379063 +0000 Modify: 2012-01-18 20:09:56.105242553 +0000 Change: 2012-01-18 20:09:56.105242553 +0000 Birth: - Similar on the other cauldron machine.
is rsyslog service started ?
You have to make sure that rsyslog.service is enabled Can you provide output of "systemctl status rsyslog.service"? (that said, I think rsyslog should have a "WantedBy=syslog.target" rather than "WantedBy=multi-user.target" which is has currently but that doesn't fully matter for the current issue).
CC: (none) => mageia
To c#5 and c#6 no syslog is not started: that is the bug. And yes it is enabled (all this is in the bugreport). In fact it is enabled since several months. But still it does not start with the last 2 reboots (equals last 2 new kernels in Cauldron). Start works, enable seems to work, but the evidence is that on reboot it is not started.
We were both talking about rsyslog.service, not syslog.service. syslog.service should be a symlink to rsyslog.service, but you've not showing anything related to the status of rsyslog.service directly (just showing status' of syslog.target above). Essentially, this symlink should exist: /etc/systemd/system/multi-user.target.wants/rsyslog.service If it's not, just type "systemctl enable rsyslog.service" and it should enable it and allow a normal boot. If not, please provide output of systemctl status rsyslog.service.
Status: NEW => ASSIGNEDAssignee: bugsquad => mageia
Colin: Sorry my bad :( Symlink exists and rebooting to kernel 3.2.2-1 the problems is fixed (although I did not change anything). So closing as resolved.
Status: ASSIGNED => RESOLVEDResolution: (none) => FIXED