since having installed patches and restarted the system, I can no longer boot into something else than the emergency shell. Trying to go further, the boot hangs just at several places, according to the messages (Applying kernel variables, Mounted /usr, …). And even CTRL-ALT-DEL does not work from there, though the keyboard works according to the LEDs, which can be switched on and off normally.
In emergency mode / and /usr are mojnted, but the status of / is
systemctl status -- -.mount
-.mount - /
Loaded: error (Reason: no such file or directory)
Active: active (mounted) since Fri, 29 Mar 2013 10:36:31 +0100; 10 minutes ago
Where: /
What: /dev/md1
CGroup: name=systemd:/system/-.mount
I tried to find out, what could be wrong in -.mount, but I did not find any file with x.mount, not even on a system, where everything was working. Later I read somewhere, that the x.mount units are generated from /etc/fstab, but that does not help with an error message like No such file or directory, if there is not file name given.
So I did a strace of systemd during a
systemctl daemon-reexec
and could systemd reading /etc/fstab and writing out the error message, but there was nothing suspicious in between.
I have no clue, what is happening.
I have no problem with /usr, it is mounted without problem. This was only to mention, that other file systems do not have a problem.
I have a problem with / or better with -.mount unter systemd, that I get this error message.
My relevant lines from /etc/fstab (/ and /usr are:
On 2013-03-29 12:56, didiw wrote:
>
> Hi,
>
> since having installed patches and restarted the system, I can no
> longer boot into something else than the emergency shell. Trying to go
> further, the boot hangs just at several places, according to the
> messages (Applying kernel variables, Mounted /usr, …). And even
> CTRL-ALT-DEL does not work from there, though the keyboard works
> according to the LEDs, which can be switched on and off normally.
> In emergency mode / and /usr are mojnted, but the status of /
> is
>
> Code:
> --------------------
> systemctl status – -.mount
> -.mount - /
> Loaded: error (Reason: no such file or directory)
> Active: active (mounted) since Fri, 29 Mar 2013 10:36:31 +0100; 10 minutes ago
> Where: /
> What: /dev/md1
> CGroup: name=systemd:/system/-.mount
> --------------------
I don’t understand why the use of systemctl to check the status of
mount. I thought that you needed only doing “mount”.
–
Cheers / Saludos,
Carlos E. R.
(from 12.1 x86_64 “Asparagus” at Telcontar)
I don’t use systemctl to check the status of the mount, but when you use systemd, it uses those units with dependencies to other units for booting. Among those units is -.mount, which seems to be the problem on my system. It is responsible for checking and mounting /, maybe also activating a raid. I have not yet found any documentation about x.mount, not where and when it is generated and what it does exactly except checking the FS and mounting.
Mount still works from the shell, but without a working -.mount unit, the system won’t boot.
–
Cheers / Saludos,
Carlos E. R.
(from 12.1 x86_64 “Asparagus” at Telcontar)
Can you verify your md1 disk is actually connected and working? - ie. that disk or disk array has power and connectivity lights on.
I don’t know what your layout is, but I suspect that it’s atypical if you’re able to boot to emergency mode without alternate media (which also suggests the possibility you have a USB disk or optical media or something else similar plugged in).
Assuming that you’re not leaving anything out of your description, I’d guess the problem is hardware related and not software…
If you’re still having a problem I’d recommend posting your entire fstab (not an excerpt) and if you can what software you may have just installed besides updating.
On 2013-03-29 16:26, didiw wrote:
>
> Hi Carlos,
>
> robin_listas;2542533 Wrote:
>> I don’t understand why the use of systemctl to check the status of
>> mount. I thought that you needed only doing “mount”.
> I don’t use systemctl to check the status of the mount, but when you
> use systemd, it uses those units with dependencies to other units for
> booting. Among those units is -.mount, which seems to be the problem on
Ah, I guess it does what “/etc/init.d/boot.rootfsck” and
“/etc/init.d/boot.localfs” did. When something there failed, you were
dumped into emergency mode. Not because there was something broken in
the scripts was broken, but because something was broken in your
system
It typically meant that a partition did not pass the filecheck, that a
partition that was listed in fstab did not appear, an error in fstab…
But those init scripts also gave appropriate information to know what
was wrong. That was systemv. Now we have systemd. They tell us it is
better. The emergency mode messages are cryptic. They told me that in
12.3 they would improve those messages. Perhaps they haven’t.
In theory you should see messages in the screen hinting at what the
problem is. There is a log where you should also see it in more detail,
but I found it impossible to decipher.
–
Cheers / Saludos,
Carlos E. R.
(from 12.1 x86_64 “Asparagus” at Telcontar)
/ seems to be correctly mounted, I can access the files on it and neither fsck or mdadm tells anything about errors.
I don’t know what your layout is, but I suspect that it’s atypical if you’re able to boot to emergency mode without alternate media (which also suggests the possibility you have a USB disk or optical media or something else similar plugged in).
It may be atypical from a newer point of view, in that I have separate boot, root and usr partitions. This is a purely historical setup. Since this is a server, all system partitions are mirrored using soft-raid (mdadm). Until 12.1 this was not a problem. And there is currently no other disk connected.
I thought that emergency mode is usually a shell used to repair partitions, that fsck could not repair itself. So it should be reachable from the normal system. Using systemd has made emergency shell available as an early shell, if I understand systemd targets correctly.
Assuming that you’re not leaving anything out of your description, I’d guess the problem is hardware related and not software…
If you’re still having a problem I’d recommend posting your entire fstab (not an excerpt) and if you can what software you may have just installed besides updating.
That is, what I think too. But it cannot be a brioken Root-fs, because the root-fs checks okay, after switching to read-only.
I also thought, that it could be something with fstab. So I removed every unnecessary space from the root-entry, but that didi not help, still the same error.
But those init scripts also gave appropriate information to know what
was wrong. That was systemv. Now we have systemd. They tell us it is
better. The emergency mode messages are cryptic. They told me that in
12.3 they would improve those messages. Perhaps they haven’t.
Being an administrator, I hate it, when developers give an error message like No such file or directory, without giving the path of the problem, so you can check on it.
This is still 12.2. Maybe I should go to 12.3.
In theory you should see messages in the screen hinting at what the
problem is. There is a log where you should also see it in more detail,
but I found it impossible to decipher.
I tried to get something out of systemd-journal, but it seems to be too early to find something useful in it. The only information I got out of dmesg.
–
Cheers / Saludos,
Carlos E. R.
(from 12.1 x86_64 “Asparagus” at Telcontar)
