GNU bug report logs -
#21349
who shows no users nowadays on Debian
Previous Next
Reported by: 積丹尼 Dan Jacobson <jidanni <at> jidanni.org>
Date: Tue, 25 Aug 2015 23:33:02 UTC
Severity: normal
Tags: notabug
Done: Assaf Gordon <assafgordon <at> gmail.com>
Bug is archived. No further changes may be made.
To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 21349 in the body.
You can then email your comments to 21349 AT debbugs.gnu.org in the normal way.
Toggle the display of automated, internal messages from the tracker.
Report forwarded
to
bug-coreutils <at> gnu.org
:
bug#21349
; Package
coreutils
.
(Tue, 25 Aug 2015 23:33:02 GMT)
Full text and
rfc822 format available.
Acknowledgement sent
to
積丹尼 Dan Jacobson <jidanni <at> jidanni.org>
:
New bug report received and forwarded. Copy sent to
bug-coreutils <at> gnu.org
.
(Tue, 25 Aug 2015 23:33:02 GMT)
Full text and
rfc822 format available.
Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):
(info "(coreutils) who invocation") says
If given no non-option arguments, ‘who’ prints the following
information for each user currently logged on: login name, terminal
line, login time, and remote hostname or X display.
Say if this means
remote hostname or remote X display.
or
X display or remote hostname.
By the way, now on Debian sid:
$ who
$ ps a
PID TTY STAT TIME COMMAND
593 tty7 Ss+ 0:10 /usr/lib/xorg/Xorg :0 -nolisten tcp vt7
1310 pts/0 Ss 0:00 bash
1338 pts/1 Ss 0:00 su -
1339 pts/1 S+ 0:00 -su
2005 tty1 Ss+ 0:00 /sbin/agetty --noclear tty1 linux
2073 pts/0 R+ 0:00 ps a
Information forwarded
to
bug-coreutils <at> gnu.org
:
bug#21349
; Package
coreutils
.
(Wed, 26 Aug 2015 01:00:05 GMT)
Full text and
rfc822 format available.
Message #8 received at 21349 <at> debbugs.gnu.org (full text, mbox):
On 26/08/15 00:14, 積丹尼 Dan Jacobson wrote:
> (info "(coreutils) who invocation") says
>
> If given no non-option arguments, ‘who’ prints the following
> information for each user currently logged on: login name, terminal
> line, login time, and remote hostname or X display.
>
> Say if this means
> remote hostname or remote X display.
>
> or
>
> X display or remote hostname.
>
> By the way, now on Debian sid:
>
> $ who
> $ ps a
> PID TTY STAT TIME COMMAND
> 593 tty7 Ss+ 0:10 /usr/lib/xorg/Xorg :0 -nolisten tcp vt7
> 1310 pts/0 Ss 0:00 bash
> 1338 pts/1 Ss 0:00 su -
> 1339 pts/1 S+ 0:00 -su
> 2005 tty1 Ss+ 0:00 /sbin/agetty --noclear tty1 linux
> 2073 pts/0 R+ 0:00 ps a
Could you look a little deeper?
who(1) calls read_utmp() from:
http://git.sv.gnu.org/gitweb/?p=gnulib.git;a=blob;f=lib/readutmp.c;hb=HEAD
thanks,
Pádraig.
Information forwarded
to
bug-coreutils <at> gnu.org
:
bug#21349
; Package
coreutils
.
(Wed, 26 Aug 2015 01:06:01 GMT)
Full text and
rfc822 format available.
Message #11 received at 21349 <at> debbugs.gnu.org (full text, mbox):
All I know is nowadays everybody (me) uses just X windows so to achieve
the who goals:
NAME
who - show who is logged on
DESCRIPTION
Print information about users who are currently logged in.
some adjustments should be made. Else it looks like there are no users
and maintenance can begin, etc.
Information forwarded
to
bug-coreutils <at> gnu.org
:
bug#21349
; Package
coreutils
.
(Wed, 26 Aug 2015 07:03:01 GMT)
Full text and
rfc822 format available.
Message #14 received at 21349 <at> debbugs.gnu.org (full text, mbox):
Hi Dan,
On Wed, Aug 26, 2015 at 07:14:41AM +0800, 積丹尼 Dan Jacobson wrote:
> (info "(coreutils) who invocation") says
>
> If given no non-option arguments, ‘who’ prints the following
> information for each user currently logged on: login name, terminal
> line, login time, and remote hostname or X display.
>
> Say if this means
> remote hostname or remote X display.
>
> or
>
> X display or remote hostname.
>
> By the way, now on Debian sid:
>
> $ who
> $ ps a
> PID TTY STAT TIME COMMAND
> 593 tty7 Ss+ 0:10 /usr/lib/xorg/Xorg :0 -nolisten tcp vt7
> 1310 pts/0 Ss 0:00 bash
> 1338 pts/1 Ss 0:00 su -
> 1339 pts/1 S+ 0:00 -su
> 2005 tty1 Ss+ 0:00 /sbin/agetty --noclear tty1 linux
> 2073 pts/0 R+ 0:00 ps a
This works on a current Debian/testing system (stable as well), so it might
be a recent Debian/Sid (unstable) issue. Perhaps you want to open a bug
report there?
Thanks,
Erik
Information forwarded
to
bug-coreutils <at> gnu.org
:
bug#21349
; Package
coreutils
.
(Wed, 26 Aug 2015 20:14:02 GMT)
Full text and
rfc822 format available.
Message #17 received at 21349 <at> debbugs.gnu.org (full text, mbox):
Erik Auerswald wrote:
> This works on a current Debian/testing system (stable as well), so it might
> be a recent Debian/Sid (unstable) issue. Perhaps you want to open a bug
> report there?
Updating utmp depends upon the terminal emulator. XTerm updates it.
Any XTerm user will be logged in utmp. I use XTerms and all of my
terminals log as a user in utmp. But most other terminal emulators
ignore utmp and don't log anything. If they don't log it then there
isn't anything 'who' can do about it.
AFAIK it doesn't have anything to do with Debian changing anything.
It is all about the changing state of Desktop Environments. For the
graphical mouse user today they have exactly one "login" now and that
is the entirety of the graphical environment and they never launch a
terminal emulator.
Bob
Information forwarded
to
bug-coreutils <at> gnu.org
:
bug#21349
; Package
coreutils
.
(Wed, 26 Aug 2015 20:32:02 GMT)
Full text and
rfc822 format available.
Message #20 received at 21349 <at> debbugs.gnu.org (full text, mbox):
On 2015-08-26 22:13 +0200, Bob Proulx wrote:
> Erik Auerswald wrote:
>> This works on a current Debian/testing system (stable as well), so it might
>> be a recent Debian/Sid (unstable) issue. Perhaps you want to open a bug
>> report there?
>
> Updating utmp depends upon the terminal emulator. XTerm updates it.
> Any XTerm user will be logged in utmp. I use XTerms and all of my
> terminals log as a user in utmp. But most other terminal emulators
> ignore utmp and don't log anything. If they don't log it then there
> isn't anything 'who' can do about it.
>
> AFAIK it doesn't have anything to do with Debian changing anything.
It most probably has, the latest xterm version (319) only writes a utmp
entry if you start a login shell (i.e. use the -ls option). That's
supposed to be a bug fix[1].
Cheers,
Sven
1. https://bugs.debian.org/794201
Information forwarded
to
bug-coreutils <at> gnu.org
:
bug#21349
; Package
coreutils
.
(Wed, 26 Aug 2015 20:41:01 GMT)
Full text and
rfc822 format available.
Message #23 received at 21349 <at> debbugs.gnu.org (full text, mbox):
Sven Joachim wrote:
> Bob Proulx wrote:
> > AFAIK it doesn't have anything to do with Debian changing anything.
>
> It most probably has, the latest xterm version (319) only writes a utmp
> entry if you start a login shell (i.e. use the -ls option). That's
> supposed to be a bug fix[1].
> 1. https://bugs.debian.org/794201
Hmm... I guess I haven't logged out since that xterm arrived.
$ who | grep -c pts/
61
... counting all of the entries up to pts/65...
And so it appears that it has been logging for me. But if I
explicitly start a new xterm:
$ xterm &
And then in that window look:
$ tty
/dev/pts/61
$ who | grep -c pts/61
0
So I stand corrected. It appears that very recently it was changed to
not log utmp. In spite of the bug log reporting the bug to have been
fixed this still seems to be active with an up to date Sid. It
appears to have arrived Sunday with version 319-1. Even explicitly
giving +ut doesn't do anything. I will further take the discussion to
the Debian bug log since it doesn't have anything to do with
coreutils.
Bob
Information forwarded
to
bug-coreutils <at> gnu.org
:
bug#21349
; Package
coreutils
.
(Thu, 27 Aug 2015 00:20:02 GMT)
Full text and
rfc822 format available.
Message #26 received at 21349 <at> debbugs.gnu.org (full text, mbox):
積丹尼 Dan Jacobson wrote:
> (info "(coreutils) who invocation") says
>
> If given no non-option arguments, ‘who’ prints the following
> information for each user currently logged on: login name, terminal
> line, login time, and remote hostname or X display.
>
> $ who
>
The same thing happens on openSuSE -- I think it's because of the move to
systemd -- it doesn't maintain the wtmp/utmp stuf..
> who
> who am i
> who -a
> who -b
> who -d
> who -H
NAME LINE TIME COMMENT
> who -l
> who -m
> who -p
> who -q
I> who -r
----
I don't have utmp, just a smallish wtmp:
So I tried a symlink:
sudo ln -s /var/log/wtmp /var/run/utmp
It's not quite the same, but some things work:
Ishtar:law> who am i
law pts/1 Aug 26 16:37 (athenae)
Ishtar:law> who
law pts/0 Aug 26 11:59 (athenae)
law pts/1 Aug 26 16:37 (athenae)
who -b doesn't work, but uptime does:
Ishtar:law> who -b
Ishtar:law> uptime
17:16pm up 61 days 12:36, 2 users, load average: 0.00, 0.06, 0.11
So has debian been moving to systemd as well?
Information forwarded
to
bug-coreutils <at> gnu.org
:
bug#21349
; Package
coreutils
.
(Thu, 27 Aug 2015 07:41:03 GMT)
Full text and
rfc822 format available.
Message #29 received at 21349 <at> debbugs.gnu.org (full text, mbox):
Hi,
On Wed, Aug 26, 2015 at 02:13:37PM -0600, Bob Proulx wrote:
> Erik Auerswald wrote:
> > This works on a current Debian/testing system (stable as well), so it might
> > be a recent Debian/Sid (unstable) issue. Perhaps you want to open a bug
> > report there?
>
> Updating utmp depends upon the terminal emulator. XTerm updates it.
Logins via X used to update it as well (seldomly used nowadays). The local
desktop session usually updates utmp as well, at least with XFCE on
Debian/testing this is still the case. GNOME Terminal updates utmp as well.
Screen updates it, too. When I last looked at it, Konsole (from KDE) did
not update utmp.
> AFAIK it doesn't have anything to do with Debian changing anything.
Sven Joachim wrote:
> It most probably has, the latest xterm version (319) only writes a utmp
> entry if you start a login shell (i.e. use the -ls option)
Linda A. Walsh noticed a similar thing:
> The same thing happens on openSuSE
Of course this is most likely caused by changes outside of coreutils. On a
"desktop" system without any terminal windows, the desktop session should
be shown in the "who" output. On all systems I could easily check that is
the case. I do not have any SystemD/GNOME or KDE systems to test.
Thanks,
Erik
Information forwarded
to
bug-coreutils <at> gnu.org
:
bug#21349
; Package
coreutils
.
(Thu, 27 Aug 2015 14:47:02 GMT)
Full text and
rfc822 format available.
Message #32 received at 21349 <at> debbugs.gnu.org (full text, mbox):
All I know is I also filed
http://bugs.debian.org/796942 w shows no users even though I must...
http://bugs.debian.org/796943 wall can't write to anybody nowadays
All I know is all this stuff should be updated to still work with the
latest Unix/Linux. Otherwise after every few years one will need xwho and
ywho and zwho just to check who's around...
Information forwarded
to
bug-coreutils <at> gnu.org
:
bug#21349
; Package
coreutils
.
(Sat, 21 Nov 2015 11:09:02 GMT)
Full text and
rfc822 format available.
Message #35 received at 21349 <at> debbugs.gnu.org (full text, mbox):
On 27/08/15 15:46, 積丹尼 Dan Jacobson wrote:
> All I know is I also filed
> http://bugs.debian.org/796942 w shows no users even though I must...
> http://bugs.debian.org/796943 wall can't write to anybody nowadays
>
> All I know is all this stuff should be updated to still work with the
> latest Unix/Linux. Otherwise after every few years one will need xwho and
> ywho and zwho just to check who's around...
A related issue with gnome-terminal on Fedora 23
https://bugzilla.redhat.com/1280724
where that doesn't update /var/run/utmp
cheers,
Pádraig
Information forwarded
to
bug-coreutils <at> gnu.org
:
bug#21349
; Package
coreutils
.
(Sun, 27 Dec 2015 21:36:01 GMT)
Full text and
rfc822 format available.
Message #38 received at 21349 <at> debbugs.gnu.org (full text, mbox):
Erik Auerswald wrote:
>
> Linda A. Walsh noticed a similar thing:
>
>> The same thing happens on openSuSE
>>
----
Looked into this a bit more and found that while
'w' and 'who' didn't work, "last" did.
Tracked it down to this...
from the linux programmer's manual UTMP(5)
it says conforming implementations should look in:
FILES
/var/run/utmp
/var/log/wtmp
--------------
tracing who (I assume 'w' uses similar code):
# strace who|& grep /var
access("/var/run/utmpx", F_OK) = -1 ENOENT (No such file or
directory)
open("/var/run/utmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or
directory)
access("/var/run/utmpx", F_OK) = -1 ENOENT (No such file or
directory)
open("/var/run/utmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or
directory)
----
I.e. it isn't checking the 2nd location.
Adding a link in the *ephemeral* (tmpfs/ramdisk) /var/run allows
it to temporarily work:
# ln -s ../log/wtmp utmp
> who
law pts/0 2015-11-29 09:49 (athenae)
root tty1 2015-12-22 09:26
law pts/1 2015-12-23 16:57 (athenae)
law pts/0 2015-12-25 17:46 (athenae)
law pts/2 2015-12-25 21:42 (athenae)
law pts/4 2015-12-25 22:23 (athenae)
law pts/5 2015-12-27 08:46 (athenae)
law pts/6 2015-12-27 12:54 (athenae)
law pts/7 2015-12-27 12:59 (athenae)
> w
13:18:22 up 108 days, 21:03, 459 users, load average: 1.33, 1.25, 0.99
USER TTY FROM LOGIN@ IDLE JCPU PCPU WHAT
law pts/0 athenae 29Nov15 39:52m 6:38 1.18s
dovecot/imap [l
root tty1 Tue09 3days 0.28s 0.28s -bash
law pts/1 athenae Wed16 2days 0.30s 0.30s -bash
law pts/0 athenae Fri17 39:52m 6:38 1.11s -bash
law pts/2 athenae Fri21 37:18m 0.42s 0.42s -bash
law pts/4 athenae Fri22 38:52m 0.25s 0.25s -bash
law pts/5 athenae 08:46 1:39m 0.57s 0.57s -bash
law pts/6 athenae 12:54 12:44 0.37s 0.37s -bash
law pts/7 athenae 12:59 0.00s 0.29s 0.01s w
> last|grep 'still log'
law pts/7 athenae Sun Dec 27 12:59 still logged in
law pts/6 athenae Sun Dec 27 12:54 still logged in
law pts/5 athenae Sun Dec 27 08:46 still logged in
law pts/4 athenae Fri Dec 25 22:23 still logged in
law pts/2 athenae Fri Dec 25 21:42 still logged in
law pts/0 athenae Fri Dec 25 17:46 still logged in
law pts/1 athenae Wed Dec 23 16:57 still logged in
root tty1 Tue Dec 22 09:26 still logged in
----------------------------------------------
While creating a symlink in /var/run/utmp => /var/log/wtmp
is a per-boot, workaround to get this to work until 'w/who' are fixed,
both of them should check the linux-standard locations documented
in the manpage.
Likely /var/log is a new location (as /var/run was made ephemeral
as part of the transition to systemd). Note, there was some
pressure by the systemd folks to also make /var/log ephemeral
as well as /tmp and /var/tmp, but those were resisted, at least
in "SUSE"...(at least for now...), since the default for systemd
is to not store or save logfiles (or coredumps) over boots.
One can tell from design choices like that systemd is designed for
end-consumer systems, where logs that document things might
be inconvenient (vs. any publicly owned corporation, law group
bank or medical office where retaining logs for some period
of time is required by US law).
Information forwarded
to
bug-coreutils <at> gnu.org
:
bug#21349
; Package
coreutils
.
(Tue, 29 Dec 2015 23:22:01 GMT)
Full text and
rfc822 format available.
Message #41 received at submit <at> debbugs.gnu.org (full text, mbox):
L. A. Walsh wrote:
> Erik Auerswald wrote:
>>
>> Linda A. Walsh noticed a similar thing:
>>
>>> The same thing happens on openSuSE
>>>
> ----
>
> Looked into this a bit more and found that while
> 'w' and 'who' didn't work, "last" did.
>
> Tracked it down to this...
> from the linux programmer's manual UTMP(5)
>
> it says conforming implementations should look in:
> FILES
> /var/run/utmp
> /var/log/wtmp
> --------------
>
> tracing who (I assume 'w' uses similar code):
> # strace who|& grep /var
> access("/var/run/utmpx", F_OK) = -1 ENOENT (No such file or
> directory)
> open("/var/run/utmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or
> directory)
> access("/var/run/utmpx", F_OK) = -1 ENOENT (No such file or
> directory)
> open("/var/run/utmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or
> directory)
> ----
>
> I.e. it isn't checking the 2nd location.
>
> Adding a link in the *ephemeral* (tmpfs/ramdisk) /var/run allows
> it to temporarily work:
>
> # ln -s ../log/wtmp utmp
>
>> who
> law pts/0 2015-11-29 09:49 (athenae)
> root tty1 2015-12-22 09:26
> law pts/1 2015-12-23 16:57 (athenae)
> law pts/0 2015-12-25 17:46 (athenae)
> law pts/2 2015-12-25 21:42 (athenae)
> law pts/4 2015-12-25 22:23 (athenae)
> law pts/5 2015-12-27 08:46 (athenae)
> law pts/6 2015-12-27 12:54 (athenae)
> law pts/7 2015-12-27 12:59 (athenae)
>> w
> 13:18:22 up 108 days, 21:03, 459 users, load average: 1.33, 1.25, 0.99
> USER TTY FROM LOGIN@ IDLE JCPU PCPU WHAT
> law pts/0 athenae 29Nov15 39:52m 6:38 1.18s
> dovecot/imap [l
> root tty1 Tue09 3days 0.28s 0.28s -bash
> law pts/1 athenae Wed16 2days 0.30s 0.30s -bash
> law pts/0 athenae Fri17 39:52m 6:38 1.11s -bash
> law pts/2 athenae Fri21 37:18m 0.42s 0.42s -bash
> law pts/4 athenae Fri22 38:52m 0.25s 0.25s -bash
> law pts/5 athenae 08:46 1:39m 0.57s 0.57s -bash
> law pts/6 athenae 12:54 12:44 0.37s 0.37s -bash
> law pts/7 athenae 12:59 0.00s 0.29s 0.01s w
>> last|grep 'still log'
> law pts/7 athenae Sun Dec 27 12:59 still logged
> in law pts/6 athenae Sun Dec 27 12:54 still
> logged in law pts/5 athenae Sun Dec 27 08:46
> still logged in law pts/4 athenae Fri Dec 25
> 22:23 still logged in law pts/2 athenae Fri
> Dec 25 21:42 still logged in law pts/0 athenae
> Fri Dec 25 17:46 still logged in law pts/1
> athenae Wed Dec 23 16:57 still logged in root
> tty1 Tue Dec 22 09:26 still logged in
>
> ----------------------------------------------
>
> While creating a symlink in /var/run/utmp => /var/log/wtmp
> is a per-boot, workaround to get this to work until 'w/who' are fixed,
> both of them should check the linux-standard locations documented
> in the manpage.
>
> Likely /var/log is a new location (as /var/run was made ephemeral
> as part of the transition to systemd). Note, there was some
> pressure by the systemd folks to also make /var/log ephemeral
> as well as /tmp and /var/tmp, but those were resisted, at least
> in "SUSE"...(at least for now...), since the default for systemd
> is to not store or save logfiles (or coredumps) over boots.
>
> One can tell from design choices like that systemd is designed for
> end-consumer systems, where logs that document things might
> be inconvenient (vs. any publicly owned corporation, law group
> bank or medical office where retaining logs for some period
> of time is required by US law).
>
>
>
Information forwarded
to
bug-coreutils <at> gnu.org
:
bug#21349
; Package
coreutils
.
(Wed, 24 Oct 2018 21:05:01 GMT)
Full text and
rfc822 format available.
Message #44 received at 21349 <at> debbugs.gnu.org (full text, mbox):
tags 21349 notabug
close 21349
stop
(triaging old bugs)
On 26/08/15 02:40 PM, Bob Proulx wrote:
> [...]
> It appears that very recently it was changed to
> not log utmp.
[...]
> I will further take the discussion to the Debian bug log since it
> doesn't have anything to do with coreutils.
With no further comments in 3 years,
I'm closing this as "not a bug" (at least not a coreutils bug).
-assaf
Added tag(s) notabug.
Request was from
Assaf Gordon <assafgordon <at> gmail.com>
to
control <at> debbugs.gnu.org
.
(Wed, 24 Oct 2018 21:05:02 GMT)
Full text and
rfc822 format available.
bug closed, send any further explanations to
21349 <at> debbugs.gnu.org and 積丹尼 Dan Jacobson <jidanni <at> jidanni.org>
Request was from
Assaf Gordon <assafgordon <at> gmail.com>
to
control <at> debbugs.gnu.org
.
(Wed, 24 Oct 2018 21:05:02 GMT)
Full text and
rfc822 format available.
Information forwarded
to
bug-coreutils <at> gnu.org
:
bug#21349
; Package
coreutils
.
(Wed, 24 Oct 2018 23:46:01 GMT)
Full text and
rfc822 format available.
Message #51 received at 21349 <at> debbugs.gnu.org (full text, mbox):
(Today on Debian I see
$ who
jidanni pts/0 2018-10-25 06:26 (:0)
So maybe this is fixed.)
bug archived.
Request was from
Debbugs Internal Request <help-debbugs <at> gnu.org>
to
internal_control <at> debbugs.gnu.org
.
(Thu, 22 Nov 2018 12:24:09 GMT)
Full text and
rfc822 format available.
This bug report was last modified 6 years and 211 days ago.
Previous Next
GNU bug tracking system
Copyright (C) 1999 Darren O. Benham,
1997,2003 nCipher Corporation Ltd,
1994-97 Ian Jackson.