[Bug 791076] New: AY installation without reboot: syslog service in failed state
https://bugzilla.novell.com/show_bug.cgi?id=791076 https://bugzilla.novell.com/show_bug.cgi?id=791076#c0 Summary: AY installation without reboot: syslog service in failed state Classification: openSUSE Product: openSUSE 12.2 Version: Final Platform: x86-64 OS/Version: openSUSE 12.2 Status: NEW Severity: Normal Priority: P5 - None Component: AutoYaST AssignedTo: fehr@suse.com ReportedBy: joschibrauchle@gmx.de QAContact: qa-bugs@suse.de Found By: --- Blocker: --- User-Agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.11 (KHTML, like Gecko) Chrome/23.0.1271.64 Safari/537.11 After performing an AY installation without a final reboot, the syslog service is in failed state: ---------- # rcsyslog status redirecting to systemctl syslog.service - System Logging Service Loaded: loaded (/lib/systemd/system/syslog.service; enabled) Active: failed (Result: exit-code) since Fri, 23 Nov 2012 17:39:07 +0100; 40min ago Process: 15328 ExecStart=/sbin/rsyslogd -c 5 -f /etc/rsyslog.conf (code=exited, status=1/FAILURE) Process: 15326 ExecStartPre=/var/run/rsyslog/addsockets (code=exited, status=0/SUCCESS) Process: 15324 ExecStartPre=/bin/systemctl stop systemd-kmsg-syslogd.service (code=exited, status=0/SUCCESS) Main PID: 804 (code=exited, status=0/SUCCESS) CGroup: name=systemd:/system/syslog.service ---------- but actually syslog (or rsyslogd in my installation) is running: ---------- #ps auxf|grep syslog root 2027 0.0 0.0 316764 1508 ? Sl 17:05 0:00 /sbin/rsyslogd -c 5 -f /etc/rsyslog.conf ---------- Possibly, syslog is started in the second installation stage by Autoyast, but after installation finished, systemd does not know about this? Starting a second instance via systemd fails, as only one instance is allowed... Reproducible: Always Steps to Reproduce: 1. Install 12.2 via AY and no final reboot 2. systemctl status syslog 3. Actual Results: syslog service failed, but rsyslogd running Expected Results: syslog service should be OK -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c1
Arvin Schnell
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c2
Frederic Crozat
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c3
Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c4
--- Comment #4 from Joschi Brauchle
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c5
--- Comment #5 from Frederic Crozat
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c6
--- Comment #6 from Frederic Crozat
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c7
--- Comment #7 from Marius Tomaschewski
Possibly, syslog is started in the second installation stage by Autoyast, but after installation finished, systemd does not know about this? Starting a second instance via systemd fails, as only one instance is allowed...
I guess, the running instance were started somehow else (manually?) or the /etc/rc.status redirection didn't redirected it, e.g. because of a start with additional arguments, e.g. "/etc/init.d/syslog start early". rsyslog has to be started via systemctl. When it has been started somehow else, it will be never accepted by systemd any more: not even for "stop". Further, when there is already a rsyslog running and you try to start another one, it will fail: xanthos:~ # /sbin/rsyslogd -c 5 -f /etc/rsyslog.conf Already running. If you want to run multiple instances, you need to specify different pid files (use -i option) Now the another way around: xanthos:~ # rcsyslog stop redirecting to systemctl xanthos:~ # /sbin/rsyslogd -c 5 -f /etc/rsyslog.conf xanthos:~ # rcsyslog start redirecting to systemctl Job failed. See system journal and 'systemctl status' for details. xanthos:~ # rcsyslog status redirecting to systemctl syslog.service - System Logging Service Loaded: loaded (/lib/systemd/system/syslog.service; enabled) Active: failed (Result: exit-code) since Mon, 26 Nov 2012 15:19:26 +0100; 2s ago Process: 11130 ExecStart=/sbin/rsyslogd -c 5 -f /etc/rsyslog.conf (code=exited, status=1/FAILURE) Process: 11128 ExecStartPre=/var/run/rsyslog/addsockets (code=exited, status=0/SUCCESS) Process: 11126 ExecStartPre=/bin/systemctl stop systemd-kmsg-syslogd.service (code=exited, status=0/SUCCESS) Main PID: 10814 (code=exited, status=0/SUCCESS) CGroup: name=systemd:/system/syslog.service # rcsyslog stop redirecting to systemctl xanthos:~ # pidof rsyslogd 11114 still running. To recover under systemd, you've to kill the running one and then start via systemd. This dirty trick works for recovery as well: xanthos:~ # pidof rsyslogd 11114 xanthos:~ # mkdir /sys/fs/cgroup/systemd/system/syslog.service/ xanthos:~ # echo 11114 > /sys/fs/cgroup/systemd/system/syslog.service/tasks xanthos:~ # rcsyslog status redirecting to systemctl syslog.service - System Logging Service Loaded: loaded (/lib/systemd/system/syslog.service; enabled) Active: failed (Result: exit-code) since Mon, 26 Nov 2012 15:19:26 +0100; 28s ago Process: 11130 ExecStart=/sbin/rsyslogd -c 5 -f /etc/rsyslog.conf (code=exited, status=1/FAILURE) Process: 11128 ExecStartPre=/var/run/rsyslog/addsockets (code=exited, status=0/SUCCESS) Process: 11126 ExecStartPre=/bin/systemctl stop systemd-kmsg-syslogd.service (code=exited, status=0/SUCCESS) Main PID: 10814 (code=exited, status=0/SUCCESS) CGroup: name=systemd:/system/syslog.service └ 11114 /sbin/rsyslogd -c 5 -f /etc/rsyslog.conf xanthos:~ # rcsyslog start redirecting to systemctl xanthos:~ # rcsyslog status redirecting to systemctl syslog.service - System Logging Service Loaded: loaded (/lib/systemd/system/syslog.service; enabled) Active: active (running) since Mon, 26 Nov 2012 15:20:01 +0100; 1s ago Process: 11159 ExecStart=/sbin/rsyslogd -c 5 -f /etc/rsyslog.conf (code=exited, status=0/SUCCESS) Process: 11157 ExecStartPre=/var/run/rsyslog/addsockets (code=exited, status=0/SUCCESS) Process: 11153 ExecStartPre=/bin/systemctl stop systemd-kmsg-syslogd.service (code=exited, status=0/SUCCESS) Main PID: 11160 (rsyslogd) CGroup: name=systemd:/system/syslog.service └ 11160 /sbin/rsyslogd -c 5 -f /etc/rsyslog.conf -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c8
--- Comment #8 from Joschi Brauchle
could you try the following : cp /lib/systemd/system/YaST2-Second-Stage.service /etc/systemd/system and in the copied file, add After=syslog.target in the [Unit] section then run "touch /var/lib/YaST2/runme_at_boot" and reboot (it will re-run the second stage)..
Did that. Actually, the system was really kinda broken after doing so (sshd not responding from outside, although enabled, nmap from another machine showed all ports closed). But as I am fighting with too many (Auto-)Yast problems already, let's ignored this for now, as it's irrelevant. The result: ------ # systemctl status syslog.service syslog.service - System Logging Service Loaded: loaded (/lib/systemd/system/syslog.service; enabled) Active: inactive (dead) since Mon, 26 Nov 2012 16:15:19 +0100; 25min ago Main PID: 810 (code=exited, status=0/SUCCESS) CGroup: name=systemd:/system/syslog.service # ps auxf|grep sysl root 1136 0.0 0.0 257460 1428 ? Sl 16:15 0:00 /sbin/rsyslogd -c 5 -f /etc/rsyslog.conf # cat /proc/1136/cgroup | grep name= 1:name=systemd:/system/YaST2-Second-Stage.service ------ syslog is still running, but not in "FAILED" state anymore, now its just "dead" after it exited at Mon, 26 Nov 2012 16:15:19 (this may be from ksyslog, as the machine booted at arount 16:15...) Just double checking: Did you really mean
cp /lib/systemd/system/YaST2-Second-Stage.service /etc/systemd/system and not /lib instead of /etc?
-- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c9
--- Comment #9 from Joschi Brauchle
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c10
--- Comment #10 from Frederic Crozat
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c11
--- Comment #11 from Joschi Brauchle
just for test, try removing SYSTEMD_NO_WRAP=1 in the .service file (there are chances a deadlock appears..)
This looks better: # systemctl status syslog.service syslog.service - System Logging Service Loaded: loaded (/lib/systemd/system/syslog.service; enabled) Active: active (running) since Mon, 26 Nov 2012 18:10:00 +0100; 6min ago Main PID: 772 (rsyslogd) CGroup: name=systemd:/system/syslog.service └ 772 /sbin/rsyslogd -c 5 -f /etc/rsyslog.conf -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c14
Jiří Suchomel
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c15
--- Comment #15 from Frederic Crozat
Since this is assigned to yast maintainers, is there anything YaST developers could do?
Could you do some tests with "serial" / "ssh" / "vnc" install, without the SYSTEMD_NO_WRAP=1 variable set, to see if we have deadlocks appearing because of this ? I'm a bit worried to change this on 12.2, without some testing in Factory first (with openQA tests) if we won't regress somewhere.. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c16
--- Comment #16 from Joschi Brauchle
I'm a bit worried to change this on 12.2, without some testing in Factory first (with openQA tests) if we won't regress somewhere..
Actually, I just did a test-installation with yast2-runlevel-2.21.5 from BNC #769924 and syslog was in "active" state afterwards... I will do more tests and report back. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c17
--- Comment #17 from Frederic Crozat
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c18
--- Comment #18 from Joschi Brauchle
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c19
Jiří Suchomel
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c20
Frederic Crozat
So what's our status here? Frederic, your pull request was already accepted, so is any change on YaST side required?
Sorry, I missed the merge (didn't receive a notification). We still need the package to be updated in Factory
Or should we close it right now and wait for new bug reports if they occur?
When aaa_base is updated in Factory, I would suggest to change YaST2-Firstboot.service and YaST2-Second-Stage.service to set Environment=SYSTEMCTL_OPTIONS=--ignore-dependencies TERM=linux instead of Environment=SYSTEMD_NO_WRAP=1 TERM=linux push the change in Factory and see if it doesn't raise any regression Rudi, could you push an updated aaa_base to Factory ? Thanks -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c21
Jiří Suchomel
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c22
Jiří Suchomel
I will update YaST2-Firstboot.service and YaST2-Second-Stage.service in yast2-installation. Could I do it right now or do I have to wait for aaa_base?
yast2-installation-2.23.7 created request id 147402 Moving to Rudi. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c23
--- Comment #23 from Bernhard Wiedemann
https://bugzilla.novell.com/show_bug.cgi?id=791076
https://bugzilla.novell.com/show_bug.cgi?id=791076#c24
Ruediger Oertel
participants (1)
-
bugzilla_noreply@novell.com