If you report a bug against libvirt, in most cases you will be asked to attach debug logs. These are bare text files which tracks transition between different states of libvirtd, what it has tried to achieve, etc. Because of client -- server schema used in libvirt, the logs can be either client or server too. Usually, it's server side that matters as nearly all interesting work takes place there. Moreover, libvirt catches stderr of all running domains. These can be useful as well.
This applies to the most common scenario of system instance of virtqemud. Log setting is not persisted, so a restart of virtqemud or the system clears this setting:
# virt-admin -c virtqemud:///system daemon-log-outputs "3:journald 1:file:/var/log/libvirt/libvirtd.log" # virt-admin -c virtqemud:///system daemon-log-filters "3:remote 4:event 3:util.json 3:util.object 3:util.dbus 3:util.netlink 3:node_device 3:rpc 3:access 1:*" # virt-admin -c virtqemud:///system daemon-timeout --timeout 0
The last command disabling timeout of the daemon is available since libvirt-8.6.0. With older versions make sure to reproduce the issue within 120 seconds or have a VM running which prevents the daemon from timing out.
For any other configuration please read the rest of the document. If you want to persist the log level and log outputs settings edit /etc/libvirt/virtqemud.conf and look for log-filters and log-outputs
Libvirt log messages are classified into 4 priority levels; the higher the priority level, the less is the volume of produced messages.
The log level setting is controlled by the log_filters and log_outputs settings explained in the Log outputs and Log filters sections respectively.
1: DEBUG
2: INFO
3: WARNING
4: ERROR
For debugging it's necessary to capture the DEBUG level entries as the name implies.
Log outputs describe where the log messages are being recorded. The outputs are described by a space-separated list of tuples in the following format:
level:output
level refers to the minimum priority level of entries recorded in the output.
output is one of the following:
- file:FILENAME
Logging messages are appended to FILENAME.
- journald
Logging goes to the journald logging daemon.
- stderr
Logging goes to the standard error output stream of the libvirt daemon.
- syslog:name
Logging goes to syslogd. name is used to identify the entries.
The default output on systems running journald is 3:journald. Note that journald can throttle the amount of logs per process so in order to capture debug logs of a libvirt daemon should go to a file instead (in addition to theoriginal logging daemon), e.g.:
"1:file:/var/log/libvirt/libvirtd.log 3:journald"
Log filters, as the name suggest, help filtering out messages which are irrelevant to the cause. The log filters is a space-separated list of tuples list of tuples using the level:identifier format. Each filter defined this way will then limit messages coming from a module matching the identifier pattern (accepts globs too) to the given level."
As identifier is based on internal naming of modules, preferred way of configuring your filters is to start with the Example filter settings.
The rule of thumb here is to have more logs rather than less and miss something important.
Libvirt daemons can be configured either via a config file or via the administration API. The configuration location depends on multiple factors.
Libvirt daemons run either in the system mode or on session (user) mode, depending on the configuration of the host and available permission levels.
The connection URI influences which daemon the client will communicate with.
all connection URIs end in /system e.g. qemu:///system
config files are usually placed in /etc/libvirt
connection URIs end in /session
config files are usually placed in $XDG_CONFIG_HOME/libvirt/ directory
While there is only a single 'libvirtd.conf' configuration file in case of the monolithic daemon setup, each of the modular daemons has their own configuration file giving you a lot of possibilities how to configure them individually including logging. Realistically though, logging will have to be configured only for a single or a couple of daemons in case debug logs are requested.
Refer to documentation about daemons to figure out which is in use by your system.
The configuration of modular daemons is in file named after the daemon. E.g. for qemu:///system connection this is the virtqemud daemon and correspondingly:
virtqemud.conf config file is used
virtqemud:///system or virtqemud:///session admin URI is used
libvirtd.conf config file is used
libvirtd:///system or libvirtd:///session admin URI is used when the modular qemu hypervisor driver virtqemud
In order to setup libvirt logging persistently, follow the steps below:
open the appropriate daemon config file in your favourite editor
/etc/libvirt/virtqemud.conf /etc/libvirt/libvirtd.conf $XDG_CONFIG_HOME/libvirt/libvirtd.conf $XDG_CONFIG_HOME/libvirt/virtqemud.conf
find & replace, or set the appropriate Log outputs and Log filters, e.g
log_filters="3:remote 4:event 3:util.json 3:rpc 1:*" log_outputs="1:file:/var/log/libvirt/libvirtd.log"
save and exit
restart the corresponding service/daemon e.g.
systemctl restart virtqemud.socket systemctl restart libvirtd.socket systemctl restart libvirtd.service
Note: Libvirt prior to the libvirt-4.4.0 release didn't support globbing patterns and thus requires more configuration. See Legacy (pre-4.4.0) libvirt daemon logging configuration.
Debugging anomalies can be very painful, especially when trying to reproduce it after the daemon restarts, since the new session can make the anomaly "disappear". Therefore, it's possible to enable the debug logs during runtime using libvirt administration API. To use it conveniently, there's the virt-admin client provided by the libvirt-admin package. Use the package manager provided by your distribution to install this package.
Important: Substitute virt-admin -c $ADMIN_URI according to the guideline in the sections above in place of virt-admin in the examples below if needed.
Common deployments of libvirt start the libvirt daemons via socket activation and with automatic shutdown timeout of 120 seconds when no client or object is handled by the daemon. When a timeout is reached the daemon stops and all settings done during runtime via virt-admin are discarded. The daemon then is re-started with another command.
To prevent auto-shutdown of the daemon you can use the following command:
virt-admin daemon-timeout --timeout 0
The above is introduced in libvirt-8.6.0.
The following command allows to query the list of currently active log filters:
# virt-admin daemon-log-filters Logging filters: 3:remote 4:util.json 4:rpc
In order to change this set, run the same command as root, this time with your own set of filters:
# virt-admin daemon-log-filters "3:remote 4:util.json 4:rpc 1:*"
Analogically, the same procedure can be performed with log outputs:
# virt-admin daemon-log-outputs Logging outputs: 3:syslog:libvirtd # virt-admin daemon-log-outputs "1:file:/var/log/libvirt/libvirtd.log"
NOTE: It's always good practice to return the settings to the original state once you're finished debugging, just remember to save the original sets of filters and outputs and restore them at the end the same way as described above.
It's also possible to remove all the filters and produce an enormous log file, but it is not recommended since some of libvirt's modules can produce a large amount of noise. However, should you really want to do this, you can specify an empty set of filters:
# virt-admin daemon-log-filters "" Logging filters:
The situation is a bit different with outputs, since libvirt always has to log somewhere and resetting the outputs to an empty set will restore the default setting which depends on the host configuration, journald in our case:
# virt-admin daemon-log-outputs Logging outputs: 1:file:/var/log/libvirt/libvirtd.log # virt-admin daemon-log-outputs "" Logging outputs: 2:journald
Old libvirt versions didn't support globbing (e.g. 1:*) to configure logging, thus it's required to explicitly set logging level to 1 (debug level) with the log_level setting and then filter out the noise with a tailored log log_filters string.
# LEGACY SETTINGS PRIOR LIBVIRT 4.4.0 log_level = 1 log_filters="1:qemu 3:remote 4:event 3:util.json 3:rpc" log_outputs="1:file:/var/log/libvirt/libvirtd.log"
Or using virt-admin:
## LEGACY APPROACH ENUMERATING ALL THE DESIRED MODULES ## # virt-admin daemon-log-filters "1:util 1:libvirt 1:storage 1:network 1:nodedev 1:qemu"
By default the client library doesn't produce any logs and usually usually it's not very interesting on its own anyway.
In case you want to get the client logs, logging is controlled via the LIBVIRT_LOG_OUTPUTS and LIBVIRT_LOG_FILTERS environment variables. Generally when client logs are needed make sure you don't filter them:
export LIBVIRT_LOG_OUTPUTS="1:file:/tmp/libvirt_client.log"
Now you should go and reproduce the bug. Once you're finished, attach:
/var/log/libvirt/libvirtd.log or whatever path you set for the daemon logs.
If the problem is related to a domain named $dom attach:
/var/log/libvirt/qemu/$dom.log (Or substitute qemu with whatever hypervisor you are using.)
The XML configuration of the vm/domain obtained by virsh dumpxml $dom
If the problem involves a crash of libvirtd or any other component, also attach the backtrace from the core dump if possible (e.g. using coredumpctl).
If you are asked for client logs, /tmp/libvirt_client.log.
Ideally don't tear down the environment in case additional information is required.
Consider whether you view any of the information in the debug logs sensitive: Sensitive information in debug logs.
Some filter setting suggestions for debugging more specific things. Unless it's explicitly stated, these work on libvirt 4.4.0 and later. Please note that some of the filters below may not log enough information for filing a proper libvirt bug. Usually it's better to log more than less.
Specifying only some sections allows for a targeted filter configuration which works on all versions and is sufficient for most cases.
1:libvirt 1:qemu 1:conf 1:security 3:event 3:json 3:file 3:object 1:util
Some subsystems are very noisy and usually not the culprit of the problems. They can be silenced individually for a less verbose log while still logging everything else. Usual suspects are the JSON code, RPC, authentication and such. A permissive filter is good for development use cases.
3:remote 4:event 3:util.json 3:util.object 3:util.dbus 3:util.netlink 3:node_device 3:rpc 3:access 1:*
This filter logs only QMP traffic and skips most of libvirt's messages.
2:qemu.qemu_monitor 3:*
Debug logs may contain information that certain users may consider sensitive although generally it's okay to share debuglogs publicly.
Information which could be deemed sensitive:
hostname of the host
names of VMs and other objects
paths to disk images
IP addresses of guests and the host
hostnames/IP addresses of disks accessed via network
Libvirt's debug logs only ever have passwords and disk encryption secrets in encrypted form without the key being part of the log. There's one notable exception, that VNC/SPICE passwords can be found in the logs.
In case you decide to mask information you consider sensitive from the posted debug logs, make sure that the masking doesn't introduce ambiguity.