[syslog-ng 3.6.1] booting with openrc, but syslog-ng thinks i boot with sytemd
Hi, First i'd like to say i have been using syslog-ng for years and am very pleased with it. Thank you for a reliable and stable product. I'm running archlinux x86_64 and boot with openrc 0.13.5 . I do have systemd installed so i can use udev functionality. Today I updated from syslog-ng 3.5.6 to 3.6.1 and found this in rc.log : [2014-12-03T13:27:21.835298] Using /dev/log Unix socket with systemd is not possible. Changing to systemd-syslog source, which supports socket activation.; * Starting syslog-ng ... [2014-12-03T13:27:21.890729] Using /dev/log Unix socket with systemd is not possible. Changing to systemd-syslog source, which supports socket activation.; [2014-12-03T13:27:21.902289] Failed to acquire /run/systemd/journal/syslog socket, disabling systemd-syslog source; [ ok ] This is the source part of my /etc/syslog-ng/syslog-ng.conf source src { unix-dgram("/dev/log"); internal(); file("/proc/kmsg"); }; I use the official archlinux syslog-ng package, https://www.archlinux.org/packages/extra/x86_64/syslog-ng/ . It uses these build flags : | ./configure--prefix=/usr--sysconfdir=/etc/syslog-ng--libexecdir=/usr/lib \ --sbindir=/usr/bin--localstatedir=/var/lib/syslog-ng--datadir=/usr/share/syslog-ng \ --with-pidfile-dir=/run--disable-spoof-source--enable-ipv6--enable-sql \ --enable-systemd--with-systemdsystemunitdir=/usr/lib/systemd/system| syslog-ng does appear to work normally, but i'd like to get rid of the incorrect error message. Lone_Wolf (full rc.log & syslog-ng.conf attached)
Hi Lone_Wolf, Please check, whether the following directory exists on your system (using openrc): /run/systemd/system/ We check this directory to determine whether systemd is used. This information can help us to solve your problem. BR, Tibor 2014-12-03 14:40 GMT+01:00 LoneVVolf <lonewolf@xs4all.nl>:
Hi,
First i'd like to say i have been using syslog-ng for years and am very pleased with it. Thank you for a reliable and stable product.
I'm running archlinux x86_64 and boot with openrc 0.13.5 . I do have systemd installed so i can use udev functionality. Today I updated from syslog-ng 3.5.6 to 3.6.1 and found this in rc.log :
[2014-12-03T13:27:21.835298] Using /dev/log Unix socket with systemd is not possible. Changing to systemd-syslog source, which supports socket activation.; * Starting syslog-ng ... [2014-12-03T13:27:21.890729] Using /dev/log Unix socket with systemd is not possible. Changing to systemd-syslog source, which supports socket activation.; [2014-12-03T13:27:21.902289] Failed to acquire /run/systemd/journal/syslog socket, disabling systemd-syslog source; [ ok ]
This is the source part of my /etc/syslog-ng/syslog-ng.conf source src { unix-dgram("/dev/log"); internal(); file("/proc/kmsg"); };
I use the official archlinux syslog-ng package, https://www.archlinux.org/packages/extra/x86_64/syslog-ng/ .
It uses these build flags :
./configure --prefix=/usr --sysconfdir=/etc/syslog-ng --libexecdir=/usr/lib \ --sbindir=/usr/bin --localstatedir=/var/lib/syslog-ng --datadir=/usr/share/syslog-ng \ --with-pidfile-dir=/run --disable-spoof-source --enable-ipv6 --enable-sql \ --enable-systemd --with-systemdsystemunitdir=/usr/lib/systemd/system
syslog-ng does appear to work normally, but i'd like to get rid of the incorrect error message.
Lone_Wolf
(full rc.log & syslog-ng.conf attached)
______________________________________________________________________________ Member info: https://lists.balabit.hu/mailman/listinfo/syslog-ng Documentation: http://www.balabit.com/support/documentation/?product=syslog-ng FAQ: http://www.balabit.com/wiki/syslog-ng-faq
On 03-12-14 17:15, Tibor Benke wrote:
Hi Lone_Wolf,
Please check, whether the following directory exists on your system (using openrc):
/run/systemd/system/
We check this directory to determine whether systemd is used. This information can help us to solve your problem.
BR, Tibor
Tibor, no /run/systemd/system/ folder present on my system. Below are the contents of my /run folder . Lone_Wolf $ ls -l --group-directories-first /run total 28 drwxr-xr-x 2 avahi avahi 80 3 dec 13:27 avahi-daemon drwxr-xr-x 2 clamav clamav 40 3 dec 13:27 clamav drwxr-xr-x 3 root lp 80 3 dec 13:27 cups drwxr-xr-x 2 root root 80 3 dec 13:27 dbus drwxr-xr-x 2 root root 40 3 dec 13:27 initramfs drwxrwxr-x 5 root uucp 120 3 dec 13:27 lock drwxr-xr-x 2 root root 40 3 dec 13:27 lvm drwxr-xr-x 2 root root 60 3 dec 13:27 mount drwxr-xr-x 2 mysql mysql 40 3 dec 13:27 mysqld drwxr-xr-x 5 root root 100 3 dec 13:27 network drwxr-xr-x 2 root root 40 3 dec 13:27 nscd drwxrwxr-x 14 root root 340 3 dec 13:27 openrc drwxr-xr-x 2 root root 40 3 dec 13:27 samba drwx--x--x 2 root root 40 3 dec 13:27 sudo drwxr-xr-x 2 root root 40 3 dec 13:27 svnserve drwxr-xr-x 2 root root 60 3 dec 13:27 tmpfiles.d drwxr-xr-x 6 root root 140 3 dec 13:27 udev drwx------ 2 root root 40 3 dec 13:28 udisks2 drwxr-xr-x 2 uuidd uuidd 40 3 dec 13:27 uuidd -rw-r--r-- 1 root root 5 3 dec 13:27 acpid.pid srw-rw-rw- 1 root root 0 3 dec 13:27 acpid.socket -rw-r--r-- 1 root root 5 3 dec 13:27 avahi-dnsconfd.pid -rw-r--r-- 1 root root 5 3 dec 13:27 crond.pid ---------- 1 root root 0 3 dec 13:27 cron.reboot -rw-r--r-- 1 root root 4 3 dec 13:27 ntpd.pid srwxr-xr-x 1 root root 0 3 dec 13:27 syslog-ng.ctl -rw-r--r-- 1 root root 5 3 dec 13:27 syslog-ng.pid -rw-rw-r-- 1 root utmp 3840 3 dec 23:17 utmp -rw-r--r-- 1 root root 5 3 dec 13:27 vsftpd.pid $
Also, you might want to use the system() source, which tries to determine the optimal setup to fetch local messages. So no need to use I unix-dgram or internal directly. On Dec 3, 2014 11:27 PM, "LoneVVolf" <lonewolf@xs4all.nl> wrote:
On 03-12-14 17:15, Tibor Benke wrote:
Hi Lone_Wolf,
Please check, whether the following directory exists on your system (using openrc):
/run/systemd/system/
We check this directory to determine whether systemd is used. This information can help us to solve your problem.
BR, Tibor
Tibor, no /run/systemd/system/ folder present on my system. Below are the contents of my /run folder . Lone_Wolf
$ ls -l --group-directories-first /run total 28 drwxr-xr-x 2 avahi avahi 80 3 dec 13:27 avahi-daemon drwxr-xr-x 2 clamav clamav 40 3 dec 13:27 clamav drwxr-xr-x 3 root lp 80 3 dec 13:27 cups drwxr-xr-x 2 root root 80 3 dec 13:27 dbus drwxr-xr-x 2 root root 40 3 dec 13:27 initramfs drwxrwxr-x 5 root uucp 120 3 dec 13:27 lock drwxr-xr-x 2 root root 40 3 dec 13:27 lvm drwxr-xr-x 2 root root 60 3 dec 13:27 mount drwxr-xr-x 2 mysql mysql 40 3 dec 13:27 mysqld drwxr-xr-x 5 root root 100 3 dec 13:27 network drwxr-xr-x 2 root root 40 3 dec 13:27 nscd drwxrwxr-x 14 root root 340 3 dec 13:27 openrc drwxr-xr-x 2 root root 40 3 dec 13:27 samba drwx--x--x 2 root root 40 3 dec 13:27 sudo drwxr-xr-x 2 root root 40 3 dec 13:27 svnserve drwxr-xr-x 2 root root 60 3 dec 13:27 tmpfiles.d drwxr-xr-x 6 root root 140 3 dec 13:27 udev drwx------ 2 root root 40 3 dec 13:28 udisks2 drwxr-xr-x 2 uuidd uuidd 40 3 dec 13:27 uuidd -rw-r--r-- 1 root root 5 3 dec 13:27 acpid.pid srw-rw-rw- 1 root root 0 3 dec 13:27 acpid.socket -rw-r--r-- 1 root root 5 3 dec 13:27 avahi-dnsconfd.pid -rw-r--r-- 1 root root 5 3 dec 13:27 crond.pid ---------- 1 root root 0 3 dec 13:27 cron.reboot -rw-r--r-- 1 root root 4 3 dec 13:27 ntpd.pid srwxr-xr-x 1 root root 0 3 dec 13:27 syslog-ng.ctl -rw-r--r-- 1 root root 5 3 dec 13:27 syslog-ng.pid -rw-rw-r-- 1 root utmp 3840 3 dec 23:17 utmp -rw-r--r-- 1 root root 5 3 dec 13:27 vsftpd.pid $
______________________________________________________________________________ Member info: https://lists.balabit.hu/mailman/listinfo/syslog-ng Documentation: http://www.balabit.com/support/documentation/?product=syslog-ng FAQ: http://www.balabit.com/wiki/syslog-ng-faq
I think you've found a bug in service-management.c: https://github.com/balabit/syslog-ng/blob/master/lib/service-management.c#L1... If you have syslog-ng compiled with --enable-systemd, it will always be the active service-management unit. I thought we addressed this in a patch by jviktor. 2014-12-04 0:04 GMT+01:00 Balazs Scheidler <bazsi77@gmail.com>:
Also, you might want to use the system() source, which tries to determine the optimal setup to fetch local messages. So no need to use I unix-dgram or internal directly. On Dec 3, 2014 11:27 PM, "LoneVVolf" <lonewolf@xs4all.nl> wrote:
On 03-12-14 17:15, Tibor Benke wrote:
Hi Lone_Wolf,
Please check, whether the following directory exists on your system (using openrc):
/run/systemd/system/
We check this directory to determine whether systemd is used. This information can help us to solve your problem.
BR, Tibor
Tibor, no /run/systemd/system/ folder present on my system. Below are the contents of my /run folder . Lone_Wolf
$ ls -l --group-directories-first /run total 28 drwxr-xr-x 2 avahi avahi 80 3 dec 13:27 avahi-daemon drwxr-xr-x 2 clamav clamav 40 3 dec 13:27 clamav drwxr-xr-x 3 root lp 80 3 dec 13:27 cups drwxr-xr-x 2 root root 80 3 dec 13:27 dbus drwxr-xr-x 2 root root 40 3 dec 13:27 initramfs drwxrwxr-x 5 root uucp 120 3 dec 13:27 lock drwxr-xr-x 2 root root 40 3 dec 13:27 lvm drwxr-xr-x 2 root root 60 3 dec 13:27 mount drwxr-xr-x 2 mysql mysql 40 3 dec 13:27 mysqld drwxr-xr-x 5 root root 100 3 dec 13:27 network drwxr-xr-x 2 root root 40 3 dec 13:27 nscd drwxrwxr-x 14 root root 340 3 dec 13:27 openrc drwxr-xr-x 2 root root 40 3 dec 13:27 samba drwx--x--x 2 root root 40 3 dec 13:27 sudo drwxr-xr-x 2 root root 40 3 dec 13:27 svnserve drwxr-xr-x 2 root root 60 3 dec 13:27 tmpfiles.d drwxr-xr-x 6 root root 140 3 dec 13:27 udev drwx------ 2 root root 40 3 dec 13:28 udisks2 drwxr-xr-x 2 uuidd uuidd 40 3 dec 13:27 uuidd -rw-r--r-- 1 root root 5 3 dec 13:27 acpid.pid srw-rw-rw- 1 root root 0 3 dec 13:27 acpid.socket -rw-r--r-- 1 root root 5 3 dec 13:27 avahi-dnsconfd.pid -rw-r--r-- 1 root root 5 3 dec 13:27 crond.pid ---------- 1 root root 0 3 dec 13:27 cron.reboot -rw-r--r-- 1 root root 4 3 dec 13:27 ntpd.pid srwxr-xr-x 1 root root 0 3 dec 13:27 syslog-ng.ctl -rw-r--r-- 1 root root 5 3 dec 13:27 syslog-ng.pid -rw-rw-r-- 1 root utmp 3840 3 dec 23:17 utmp -rw-r--r-- 1 root root 5 3 dec 13:27 vsftpd.pid $
______________________________________________________________________________ Member info: https://lists.balabit.hu/mailman/listinfo/syslog-ng Documentation: http://www.balabit.com/support/documentation/?product=syslog-ng FAQ: http://www.balabit.com/wiki/syslog-ng-faq
______________________________________________________________________________ Member info: https://lists.balabit.hu/mailman/listinfo/syslog-ng Documentation: http://www.balabit.com/support/documentation/?product=syslog-ng FAQ: http://www.balabit.com/wiki/syslog-ng-faq
On 04-12-14 09:58, Tibor Benke wrote:
I think you've found a bug in service-management.c:
https://github.com/balabit/syslog-ng/blob/master/lib/service-management.c#L1...
If you have syslog-ng compiled with --enable-systemd, it will always be the active service-management unit. I thought we addressed this in a patch by jviktor. Tibor,
Is that code run during compiling/building or at run-time ? Archlinux developers build packages on machines with an archlinux base installation with systemd installed & active . If service-management.c is executed at build-time, that would explain why syslog-ng tries to use systemd on my system. Given the multi-platform support of syslog-ng i would expect such checks to be executed at run-time .
2014-12-04 0:04 GMT+01:00 Balazs Scheidler <bazsi77@gmail.com <mailto:bazsi77@gmail.com>>:
Also, you might want to use the system() source, which tries to determine the optimal setup to fetch local messages. So no need to use I unix-dgram or internal directly.
Balazs, thanks for the tip. After the problem is fixed, i'll look into using that.
Hi, Viktor could you pls share the fix for this? I know I promised to port it, but got distracted. Thanks. On Dec 4, 2014 11:36 AM, "LoneVVolf" <lonewolf@xs4all.nl> wrote:
On 04-12-14 09:58, Tibor Benke wrote:
I think you've found a bug in service-management.c:
https://github.com/balabit/syslog-ng/blob/master/lib/service-management.c#L1...
If you have syslog-ng compiled with --enable-systemd, it will always be the active service-management unit. I thought we addressed this in a patch by jviktor.
Tibor,
Is that code run during compiling/building or at run-time ? Archlinux developers build packages on machines with an archlinux base installation with systemd installed & active . If service-management.c is executed at build-time, that would explain why syslog-ng tries to use systemd on my system.
Given the multi-platform support of syslog-ng i would expect such checks to be executed at run-time .
2014-12-04 0:04 GMT+01:00 Balazs Scheidler <bazsi77@gmail.com>:
Also, you might want to use the system() source, which tries to determine the optimal setup to fetch local messages. So no need to use I unix-dgram or internal directly.
Balazs, thanks for the tip. After the problem is fixed, i'll look into using that.
______________________________________________________________________________ Member info: https://lists.balabit.hu/mailman/listinfo/syslog-ng Documentation: http://www.balabit.com/support/documentation/?product=syslog-ng FAQ: http://www.balabit.com/wiki/syslog-ng-faq
participants (3)
-
Balazs Scheidler
-
LoneVVolf
-
Tibor Benke