How to debug Systemd problems

From FedoraProject

(Difference between revisions)
Jump to: navigation, search
(shutdown problems)
(reference the upstream wiki rather than maintaining a stale copy of it)
 
(3 intermediate revisions by one user not shown)
Line 5: Line 5:
 
If you are experiencing a problem with system boot up due to [[Systemd]], please see the [[Bugs/Common|common bugs]] document before filing a bug.  Some easy configuration tweaks that fix a wide range of issues may be listed there.  If the problem you are seeing is not listed there or none of the workarounds seem to help, please consider filing a bug to help us make Fedora run better on your hardware.
 
If you are experiencing a problem with system boot up due to [[Systemd]], please see the [[Bugs/Common|common bugs]] document before filing a bug.  Some easy configuration tweaks that fix a wide range of issues may be listed there.  If the problem you are seeing is not listed there or none of the workarounds seem to help, please consider filing a bug to help us make Fedora run better on your hardware.
  
Be prepared to include some information (logs) about your system as well. These should be complete (no snippets please), not in an archive, uncompressed, with MIME type set as text/plain.
 
  
= Diagnosing boot problems =
+
= Debugging systemd problems =
If your machine gets stuck during boot, first check if the hang happens before or after control passes to systemd.
+
{{admon/note|Follow the upstream wiki guide|To debug systemd problems please follow the helpful upstream wiki page on the topic:<br>
 +
[http://freedesktop.org/wiki/Software/systemd/Debugging http://freedesktop.org/wiki/Software/systemd/Debugging]
 +
}}
  
Try to boot without <code>rhgb</code> and <code>quiet</code> on the kernel command line.
+
= Various useful systemd related commands =
If you see some messages like these:
+
* Welcome to <span style="color:blue">Fedora ''VERSION'' (''codename'')</span>!"
+
* Starting ''name''...
+
* Started ''name''. [  <span style="color:green">OK</span>  ]
+
then systemd is running.
+
 
+
{{admon/note|Problems booting this far?|If you experience bootup difficulties before systemd starts running, see [[How_to_debug_Dracut_problems|Debugging Dracut]] to diagnose that further. }}
+
 
+
{{admon/note|Problems running udev?|If you're system hangs while starting or running udev, see [[How_to_debug_Udev_problems|Debugging Udev]] to diagnose that further. }}
+
 
+
Debugging always gets easier when you can get a shell. If you do not get a login prompt, try switching to a different virtual terminal using CTRL+ALT+F''<number>''. Problems with X server startup may manifest themselves as a missing login on tty1, but other VTs working.
+
 
+
If the boot stops without presenting you with a login on any virtual console, give it '''up to 5 minutes''' before declaring it definitely stuck. There is a chance that a service that has trouble starting will be killed after this timeout and the boot will continue normally. Another possibility is that a device for an important mountpoint will fail to appear and you will be presented with '''emergency mode'''.
+
 
+
== If you get no shell ==
+
If you get neither a normal login nor the emergency mode shell, you will need to do additional steps to get debugging information out of the machine.
+
* Try CTRL+ALT+DEL to reboot.
+
** If it does not reboot, mention it in your bugreport. Meanwhile force the reboot with SysRq or hard reset.
+
* When booting the next time, you will have to add some kernel command line arguments depending on what debugging strategy you choose from the following options.
+
 
+
=== strategy: debug logging to a serial console ===
+
If you have a hardware serial console available or if you are debugging in a virtual machine (e.g. using virt-manager you can switch your view to a serial console in the menu View -> Text Consoles), you can ask systemd to log lots of useful debugging information to it by booting with:
+
<pre>
+
systemd.log_level=debug systemd.log_target=console console=ttyS0,38400
+
</pre>
+
 
+
=== strategy: booting into rescue or emergency targets ===
+
To boot directly into rescue target add <code>systemd.unit=rescue.target</code> or <code>1</code> to the kernel command line. This target is useful if the problem occurs somewhere after the basic system is brought up, during the starting of "normal" services. If this is the case, you should be able to disable the bad service from here. If the rescue target will not boot either, the more minimal emergency target might.
+
 
+
To boot directly into emergency shell add <code>systemd.unit=emergency.target</code> or <code>emergency</code> to the kernel command line. Note that in the emergency shell you will have to remount the root filesystem read-write by yourself before editing any files:
+
<pre>
+
mount -o remount,rw /
+
</pre>
+
Common issues that can be resolved in the emergency shell are bad lines in <code>/etc/fstab</code>. After fixing <code>/etc/fstab</code>, run <code>systemctl daemon-reload</code> to let systemd refresh its view of it.
+
 
+
If not even the emergency target works, you can boot directly into a shell with <code>init=/bin/sh</code>. This may be necessary in case systemd itself or some libraries it depends on are damaged by filesystem corruption. You may need to reinstall working versions of the affected packages.
+
 
+
If <code>/bin/sh</code> does not work, you must boot from another medium.
+
 
+
=== strategy: early debug shell ===
+
You can enable shell access to be available very early in the startup process to fall back on and diagnose systemd related boot up issues with various systemctl commands.
+
 
+
* Follow the instructions for [[systemd_early_debug_shell|enabling an early debug shell]].
+
 
+
== When you can get a shell ==
+
When you have systemd running to the extent that it can provide you with a shell, please use it to extract useful information for debugging. Boot with these parameters on the kernel command line:
+
<pre>systemd.log_level=debug systemd.log_target=kmsg log_buf_len=1M</pre>
+
in order to increase the verbosity of systemd, to let systemd write its logs to the kernel log buffer, and to increase the size of the kernel log buffer. After reaching the shell, save the log:
+
<pre>
+
dmesg > dmesg.txt
+
</pre>
+
 
+
When reporting a bug, attach the <code>dmesg.txt</code> file.
+
 
+
To check for possibly stuck jobs use:
+
<pre>
+
systemctl list-jobs
+
</pre>
+
 
+
A useful command to run is this, which test-runs systemd and calculates the initial transaction:
+
<pre>
+
/lib/systemd/systemd --test --system --log-level=debug
+
</pre>
+
 
+
= Diagnosing shutdown problems =
+
If reboot or poweroff take suspiciously long time,
+
* boot with the debug options:
+
<pre>systemd.log_level=debug systemd.log_target=kmsg log_buf_len=1M enforcing=0</pre>
+
* save the following script as '''/lib/systemd/system-shutdown/debugshutdown.sh''' and make it executable:
+
<pre>
+
#!/bin/sh
+
mount -o remount,rw /
+
dmesg > /shutdown-log.txt
+
mount -o remount,ro /
+
</pre>
+
* reboot
+
Look for timeouts logged in the resulting file <code>/shutdown-log.txt</code> and/or attach it to a bugreport.
+
 
+
= Information to include in your report =
+
 
+
{{Anchor|AllInfo}}
+
 
+
== All bug reports ==
+
 
+
In all cases, the following should be mentioned and attached to your bug report:
+
 
+
* The exact kernel command-line used if not default.  Typically from the bootloader configuration file (e.g. {{filename|/etc/grub.conf}}) or from {{filename|/proc/cmdline}}
+
* A copy of the file {{filename|/var/log/messages}}
+
* The output of the dmesg command: <code>dmesg > dmesg.txt</code>
+
* The output of a systemd dump: <code>systemctl dump > systemd-dump.txt</code>
+
* The output of <code>/bin/systemd --test --system --log-level=debug > systemd-test.txt 2>&1</code>
+
 
+
== Configure a serial console ==
+
 
+
Successfully debugging boot up will require some form of console logging during the system boot.  This section documents configuring a serial console connection to record boot messages.
+
 
+
First, enable serial console output for both the kernel and the bootloader.
+
Open the file {{filename|/etc/grub.conf}} for editing.  Below the line ''timeout=5'', add the following:
+
<pre>
+
serial --unit=0 --speed=38400
+
terminal --timeout=5 serial console
+
</pre>
+
Also in {{filename|/etc/grub.conf}}, add the following boot arguments to the ''kernel'' line:
+
<pre>
+
console=tty0 console=ttyS0,38400
+
</pre>
+
When finished, the {{filename|/etc/grub.conf}} file should look similar to the example below.
+
<pre>
+
default=0
+
timeout=5
+
serial --unit=0 --speed=38400
+
terminal --timeout=5 serial console
+
title Fedora (2.6.35.4-18.fc14.x86_64)
+
        root (hd0,0)
+
        kernel /vmlinuz-2.6.35.4-18.fc14.x86_64 ro root=/dev/mapper/VolGroup-lv_root rd_LVM_LV=VolGroup/lv_root rd_LVM_LV=VolGroup/lv_swap rd_NO_LUKS rd_NO_MD rd_NO_DM LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 KEYBOARDTYPE=pc KEYTABLE=is-latin1 systemd.log_level=debug systemd.log_target=kmsg console=tty0 console=ttyS0,38400
+
        initrd /initramfs-2.6.35.4-18.fc14.x86_64.img
+
</pre>
+
More detailed information on how to configure the kernel for console output can be found at [http://www.faqs.org/docs/Linux-HOWTO/Remote-Serial-Console-HOWTO.html#CONFIGURE-KERNEL].
+
 
+
{{admon/tip|Redirecting non-interactive output|You can redirect all non-interactive output to /dev/kmsg and the kernel will put it out on the console when it reaches the kernel buffer by doing
+
<pre>exec >/dev/kmsg 2>&1 </dev/console</pre>}}
+
 
+
== Boot into rescue mode or to a emergency shell ==
+
 
+
To boot directly into rescue mode add <code>systemd.unit=rescue.target</code> or <code>1</code> to the kernel command line.
+
 
+
To boot directly into emergency shell add <code>systemd.unit=emergency.target</code> or <code>emergency</code> to the kernel command line.
+
 
+
== Enable shell access early in systemd ==
+
 
+
You can enable shell access early in systemd startup process to fall back on and diagnose systemd related boot up issues with various systemctl commands should systemd startup otherwise fail.
+
 
+
* Follow the instructions for [[systemd_early_debug_shell|enabling an early debug shell]].
+
 
+
== Various useful systemd related commands ==
+
  
 
* Run <code>systemctl list-jobs</code>
 
* Run <code>systemctl list-jobs</code>
  
 
To identify slow boot and look for the jobs that are "running" those jobs are the ones where boot waits for completion on and the ones that listed as "waiting" will be executed only after those which are "running" are completed.
 
To identify slow boot and look for the jobs that are "running" those jobs are the ones where boot waits for completion on and the ones that listed as "waiting" will be executed only after those which are "running" are completed.
 
  
 
* Run <code>systemctl list-units -t service --all</code>  
 
* Run <code>systemctl list-units -t service --all</code>  
Line 186: Line 51:
 
To examine what gets started when when booted into a specific target. ( In the above example the multi-user.target )
 
To examine what gets started when when booted into a specific target. ( In the above example the multi-user.target )
  
== Systemd boot parameters ==
+
= Systemd boot parameters =
  
 
The following boot parameters are also available to further assist with debugging boot issues.
 
The following boot parameters are also available to further assist with debugging boot issues.

Latest revision as of 12:45, 15 June 2012


Foreword

If you are experiencing a problem with system boot up due to Systemd, please see the common bugs document before filing a bug. Some easy configuration tweaks that fix a wide range of issues may be listed there. If the problem you are seeing is not listed there or none of the workarounds seem to help, please consider filing a bug to help us make Fedora run better on your hardware.


[edit] Debugging systemd problems

Note.png
Follow the upstream wiki guide
To debug systemd problems please follow the helpful upstream wiki page on the topic:
http://freedesktop.org/wiki/Software/systemd/Debugging

[edit] Various useful systemd related commands

  • Run systemctl list-jobs

To identify slow boot and look for the jobs that are "running" those jobs are the ones where boot waits for completion on and the ones that listed as "waiting" will be executed only after those which are "running" are completed.

  • Run systemctl list-units -t service --all

To list all available services and their current status


  • Run systemctl list-units -t service

To show all active services


  • Run systemctl status sshd.service

To examine the current runtime status of a service. (In the above example the ssh service)


  • Run systemctl list-units -t target --all

To show all available targets.


  • Run systemctl list-units -t target

To show all active targets.


  • Run systemctl show -p "Wants" multi-user.target

To see which services a target pulls in. ( In the above example the multi-user.target )


  • Run systemd --test --system --unit=multi-user.target

To examine what gets started when when booted into a specific target. ( In the above example the multi-user.target )

[edit] Systemd boot parameters

The following boot parameters are also available to further assist with debugging boot issues.

systemd.unit= 
Overrides the unit to activate on boot. This may be used to temporarily boot into a different boot unit, for example rescue.target or emergency.target. ( Defaults to default.target. )
systemd.dump_core= 
Takes a boolean argument. If true systemd dumps core when it crashes. Otherwise no core dump is created. ( Defaults to true )
systemd.crash_shell= 
Takes a boolean argument. If true systemd spawns a shell when it crashes. Otherwise no core dump is created. Defaults to false, for security reasons, as the shell is not protected by any password authentication.
systemd.crash_chvt= 
Takes an integer argument. If positive systemd activates the specified virtual terminal when it crashes. ( Defaults to -1 )
systemd.confirm_spawn= 
Takes a boolean argument. If true asks for confirmation when spawning processes. ( Defaults to false )
systemd.show_status= 
Takes a boolean argument. If true shows terse service status updates on the console during bootup. ( Defaults to true )
systemd.sysv_console= 
Takes a boolean argument. If true output of SysV init scripts will be directed to the console. ( Defaults to true, unless quiet is passed as kernel command line option in which case it defaults to false. )
systemd.log_target= 
Set log target. Argument must be one of console, syslog, kmsg, syslog-or-kmsg, null.
systemd.log_level= 
Set log level. As argument this accepts a numerical log level or the well-known syslog symbolic names (lowercase): emerg, alert, crit, err, warning, notice, info, debug.
systemd.log_color= 
Highlight important log messages. Argument is a boolean value. If the argument is omitted it defaults to true.
systemd.log_location= 
Include code location in log messages. This is mostly relevant for debugging purposes. Argument is a boolean value. If the argument is omitted it defaults to true.