5.166. libvirt

Updated libvirt packages that fix several bugs are now available for Red Hat Enterprise Linux 6.
The libvirt library is a C API for managing and interacting with the virtualization capabilities of Linux and other operating systems. In addition, libvirt provides tools for remote management of virtualized systems.

Bug Fixes

BZ#877024
The AMD FX series processors contain "modules" which are reported by the kernel as both threads and cores. Previously, the processor topology detection code in libvirt was unable to detect these modules. Consequently, libvirt reported twice the actual number of processors. With this update, topologies that add up to the total number of processors reported by the system are properly reported even though the actual topology has to be checked in the output of the virCapabilities() function.
Note that the capability output for topology detection purposes should be used due to performance reasons. The NUMA topology has high impact on performance but the impact of the physical topology can differ from that.
BZ#884713
Whenever the virDomainGetXMLDesc() function was executed on a domain that was unresponsive, the call also became unresponsive. With this update, QEMU sends the BALLOON_CHANGE event when memory usage on a domain changes so that virDomainGetXMLDesc() no longer has to query an unresponsive domain. As a result, virDomainGetXMLDesc() calls no longer hang in the described scenario.
All users of libvirt are advised to upgrade to these updated packages, which fix these bugs.
Updated libvirt packages that fix one security issue are now available for Red Hat Enterprise Linux 6.
The Red Hat Security Response Team has rated this update as having important security impact. A Common Vulnerability Scoring System (CVSS) base score, which gives a detailed severity rating, is available from the CVE link(s) associated with each description below.
The libvirt library is a C API for managing and interacting with the virtualization capabilities of Linux and other operating systems. In addition, libvirt provides tools for remote management of virtualized systems.

Security Fix

CVE-2013-0170
A flaw was found in the way libvirtd handled connection cleanup (when a connection was being closed) under certain error conditions. A remote attacker able to establish a read-only connection to libvirtd could use this flaw to crash libvirtd or, potentially, execute arbitrary code with the privileges of the root user.
This issue was discovered by Tingting Zheng of Red Hat.
All users of libvirt are advised to upgrade to these updated packages, which contain a backported patch to correct this issue. After installing the updated packages, libvirtd will be restarted automatically.
Updated libvirt packages that fix one security issue and multiple bugs are now available for Red Hat Enterprise Linux 6.
The Red Hat Security Response Team has rated this update as having moderate security impact. A Common Vulnerability Scoring System (CVSS) base score, which gives a detailed severity rating, is available for each vulnerability from the CVE link(s) associated with each description below.
The libvirt library is a C API for managing and interacting with the virtualization capabilities of Linux and other operating systems. In addition, libvirt provides tools for remote management of virtualized systems.

Security Fix

CVE-2012-4423
A flaw was found in libvirtd's RPC call handling. An attacker able to establish a read-only connection to libvirtd could use this flaw to crash libvirtd by sending an RPC message that has an event as the RPC number, or an RPC number that falls into a gap in the RPC dispatch table.
This issue was discovered by Wenlong Huang of the Red Hat Virtualization QE Team.

Bug Fixes

BZ#858988
When the host_uuid option was present in the libvirtd.conf file, the augeas libvirt lens was unable to parse the file. This bug has been fixed and the augeas libvirt lens now parses libvirtd.conf as expected in the described scenario.
BZ#859376
Disk hot plug is a two-part action: the qemuMonitorAddDrive() call is followed by the qemuMonitorAddDevice() call. When the first part succeeded but the second one failed, libvirt failed to roll back the first part and the device remained in use even though the disk hot plug failed. With this update, the rollback for the drive addition is properly performed in the described scenario and disk hot plug now works as expected.
BZ#860720
When a virtual machine was started with an image chain using block devices and a block rebase operation was issued, the operation failed on completion in the blockJobAbort() function. This update relabels and configures cgroups for the backing files and the rebase operation now succeeds.
All users of libvirt are advised to upgrade to these updated packages, which contain backported patches to correct these issues. After installing the updated packages, libvirtd will be restarted automatically.
Updated libvirt packages that fix one security issue and two bugs are now available for Red Hat Enterprise Linux 6.
The Red Hat Security Response Team has rated this update as having moderate security impact. A Common Vulnerability Scoring System (CVSS) base score, which gives a detailed severity rating, is available for each vulnerability from the CVE link(s) associated with each description below.
The libvirt library is a C API for managing and interacting with the virtualization capabilities of Linux and other operating systems. In addition, libvirt provides tools for remote management of virtualized systems.

Security Fix

CVE-2012-3445
A flaw was found in libvirtd's RPC call handling. An attacker able to establish a read-only connection to libvirtd could trigger this flaw with a specially-crafted RPC command that has the number of parameters set to 0, causing libvirtd to access invalid memory and crash.

Bug Fixes

BZ#847946
Previously, repeatedly migrating a guest between two machines while using the tunnelled migration could cause the libvirt daemon to lock up unexpectedly. The bug in the code for locking remote drivers has been fixed and repeated tunnelled migrations of domains now work as expected.
BZ#847959
Previously, when certain system locales were used by the system, libvirt could issue incorrect commands to the hypervisor. This bug has been fixed and the libvirt library and daemon are no longer affected by the choice of the user locale.
All users of libvirt are advised to upgrade to these updated packages, which contain backported patches to correct these issues. After installing the updated packages, libvirtd will be restarted automatically.
Updated libvirt packages that fix several bugs are now available for Red Hat Enterprise Linux 6.
The libvirt library is a C API for managing and interacting with the virtualization capabilities of Linux and other operating systems. In addition, libvirt provides tools for remote management of virtualized systems.

Bug Fixes

BZ#868972
When libvirt could not find a suitable CPU model for a host CPU, it would not provide the CPU topology in host capabilities even though the topology was detected correctly. Consequently, applications that work with the host CPU topology but not with the CPU model could not see the topology in host capabilities. With this update, the host capabilities XML description contains the host CPU topology even if the host CPU model is unknown.
BZ#869650
Previously, the fixed limit for the maximum size of an RPC message that could be supplied to the libvirtd daemon (65536 bytes) was not always sufficient. Consequently, messages that were longer than that could be dropped, leaving a client unable to fetch important data. With this update, the buffer for incoming messages has been made dynamic and libvirtd now allocates as much memory as is needed for a given message, thus allowing to send much bigger messages.
BZ#869723
Prior to this update, libvirt used an unsuitable detection procedure to detect NUMA and processor topology of a system. Consequently, topology of some advanced multi-processor systems was detected incorrectly and management applications could not utilize the full potential of the system. Now, the detection has been improved and the topology is properly recognized even on modern systems.
BZ#873292
Under certain circumstances, the iohelper process failed to write data to disk while saving a domain and kernel did not report an out-of-space error (ENOSPC). With this update, libvirt calls the fdatasync() function in the described scenario to force the data to be written to disk or catch a write error. As a result, if a write error occurs, it is now properly caught and reported.
BZ#874235
Certain operations in libvirt can be done only when a domain is paused to prevent data corruption. However, if a resuming operation failed, the management application was not notified since no event was sent. This update introduces the VIR_DOMAIN_EVENT_SUSPENDED_API_ERROR event and management applications can now keep closer track of domain states and act accordingly.
BZ#875770
Libvirt allows users to cancel an ongoing migration. Previously, if an attempt to cancel the migration was made in the migration preparation phase, qemu missed the request and the migration was not canceled. With this update, the virDomainAbortJob() function sets a flag when a cancel request is made and this flag is checked before the main phase of the migration starts. As a result, a migration can now be properly canceled even in the preparation phase.
BZ#875788
When a qemu process is being destroyed by libvirt, a clean-up operation frees some internal structures and locks. However, since users can destroy qemu processes at the same time, libvirt holds the qemu driver lock to protect the list of domains and their states, among other things. Previously, a function tried to set up the qemu driver lock when it was already up, creating a deadlock. The code has been modified to always check if the lock is free before attempting to set it up, thus fixing this bug.
All users of libvirt are advised to upgrade to these updated packages, which fix these bugs.
Updated libvirt packages that fix two bugs are now available for Red Hat Enterprise Linux 6.
The libvirt library is a C API for managing and interacting with the virtualization capabilities of Linux and other operating systems. In addition, libvirt provides tools for remote management of virtualized systems.

Bug Fixes

BZ#836916
Previously, repeatedly attaching and detaching a PCI device to a guest domain could cause the libvirt daemon to terminate unexpectedly. The erroneous structure free operation at the root of this bug has been fixed and repeated attach and detach actions of a PCI device now work as expected.
BZ#836919
On certain NUMA architectures, libvirt was failing to process and expose the NUMA topology, possibly leading to performance degradation. These updated packages now correctly parse and expose the NUMA topology on such machines and make the correct CPU placement, thus avoiding the performance degradation.
BZ#838819
When using the sanlock daemon for locking resources used by a domain, if such a resource was read-only, the locking attempt failed. Consequently, it was impossible to start a domain with a CD-ROM drive. This bug has been fixed and sanlock can now be properly used with read-only devices.
All users of libvirt are advised to upgrade to these updated packages, which fix these bugs.
Updated libvirt packages that fix two bugs are now available for Red Hat Enterprise Linux 6.
The libvirt library is a C API for managing and interacting with the virtualization capabilities of Linux and other operating systems. In addition, libvirt provides tools for remote management of virtualized systems.

Bug Fixes

BZ#827050
Closing a file descriptor multiple times could, under certain circumstances, lead to a failure to execute the qemu-kvm binary. As a consequence, a guest failed to start. A patch has been applied to address this issue, so that the guest now starts successfully.
BZ#832184
Libvirt 0.9.10 has added support for keepalive checking to detect broken connections between the client and the server. However, due to bugs in the implementation this could have caused a failure of service and disconnection, for example, during parallel migrations. The keepalive support is now disabled by default and random disconnections no longer occur.
All users of libvirt are advised to upgrade to these updated packages, which fix these bugs.
Updated libvirt packages that fix one security issue, multiple bugs, and add various enhancements are now available for Red Hat Enterprise Linux 6.
The Red Hat Security Response Team has rated this update as having low security impact. A Common Vulnerability Scoring System (CVSS) base score, which gives a detailed severity rating, is available from the CVE link associated with the description below.
The libvirt library is a C API for managing and interacting with the virtualization capabilities of Linux and other operating systems. In addition, libvirt provides tools for remote management of virtualized systems.

Note

The libvirt packages have been upgraded to upstream version 0.9.10, which provides a number of bug fixes and enhancements over the previous version. (BZ#752433)

Security Fix

CVE-2012-2693
Bus and device IDs were ignored when attempting to attach multiple USB devices with identical vendor or product IDs to a guest. This could result in the wrong device being attached to a guest, giving that guest root access to the device.

Bug Fixes

BZ#754621
Previously, libvirt incorrectly released resources in the macvtap network driver in the underlying code for QEMU. As a consequence, after an attempt to create a virtual machine failed, a macvtap device that was created for the machine could not be deleted from the system. Any virtual machine using the same MAC address could not be created in such a case. With this update, an incorrect function call has been removed, and macvtap devices are properly removed from the system in the scenario described.
BZ#742087
Under certain circumstances, a race condition between asynchronous jobs and query jobs could occur in the QEMU monitor. Consequently, after the QEMU guest was stopped, it failed to start again with the following error message:
error: Failed to start domain [domain name]
error: Timed out during operation cannot acquire state change lock
With this update, libvirt handles this situation properly, and guests now start as expected.
BZ#769500
Previously, libvirt defined a hard limit for the maximum number of virtual machines (500) in Python bindings. As a consequence, the vdsmd daemon was unable to properly discover all virtual machines on a system with more than 500 guests. With this update, the number of virtual machines is now determined dynamically and vdsmd correctly discovers all virtual machines.
BZ#739075
Previously, it was not possible to cancel all migration-family commands (for example, it was possible to cancel the "virsh migration" command, but not the "virsh dump" command). This update implements a mechanism used for "virsh migration" also for the other commands, so it is now possible to cancel these commands.
BZ#773667
Previously, libvirt was unable to verify if there were multiple active PCI devices on the same I/O bus. As a consequence, the "virsh attach-device" command failed even if such a device had already been detached from the host. With this update, libvirt properly checks for active devices on the same PCI I/O bus. Users can now attach devices to a guest successfully if the devices on the same bus are detached from the host.
BZ#701654
When the libvirt's virDomainDestroy API is shutting down the qemu process, the API first sends the SIGTERM signal, then waits for 1.6 seconds and, if the process is still running, the API sends the SIGKILL signal. Previously, this could lead to data loss because the guest running in QEMU did not have time to flush its disk cache buffers before it was unexpectedly killed. This update provides a new flag, "VIR_DOMAIN_DESTROY_GRACEFUL". If this flag is set in the call to virDomainDestroyFlags, SIGKILL is not sent to the qemu process; instead, if the timeout is reached and the qemu process still exists, virDomainDestroy returns an error. It is recommended that management applications always first call virDomainDestroyFlags with VIR_DOMAIN_DESTROY_GRACEFUL. If that fails, then the application can decide if and when to call virDomainDestroyFlags again without VIR_DOMAIN_DESTROY_GRACEFUL.
BZ#784151
The localtime_r() function used in the libvirt code was not async-signal safe, which caused child processes to enter a deadlock when attempting to generate a log message. As a consequence, the virsh utility became unresponsive. This update applies backported patches and adds a new API for generating log time stamps in an async-signal safe manner. The virsh utility no longer hangs under these circumstances.
BZ#785269
Previously, if the libvirt package was built with Avahi support, libvirt required the avahi package to be installed on the system as a prerequisite for its own installation. If the avahi package could not be installed on the system due to security concerns, installation of libvirt failed. This update modifies the libvirt.spec file to require only the avahi-libs package. The libvirt package is now successfully installed and libvirtd starts as expected.
BZ#639599
The schema for the XML files contained stricter rules than those that were actually enforced by libvirt. As a consequence, validation tools failed to validate guest XML files that contained special characters in the guest's name even if libvirt accepted the XML file. With this update, the XML schema now allows arbitrary strings with no limitation, leaving the enforcement of rules to the hypervisor driver. As a result, users are now able to validate these XML files.
BZ#785164
Previously, the libxml2 tool did not parse IPv6 URIs as expected. As a consequence, attempting to establish an IPv6 connection through SSH failed, because an invalid IPv6 address was used. A patch has been applied to address this problem and IPv6 connections can now be established successfully in this scenario.
BZ#625362
Previously, the libvirt-guests init script executed operations on guests serially. Consequently, on machines with many guests, the shutdown process took a long time because guests were waiting for other guests to be shut down. The libvirt-guests init script was modified to enable parallel operation on domains, which reduces the time of the shutdown process of the host. Now, guests start and shut down in parallel, and utilize the host system's resources more efficiently.
BZ#783968
When migrating a QEMU virtual machine and using SPICE for a remote display, the migration was failing and the display was erratic under certain circumstances. This was happening because with the incoming migration connection open, QEMU was unable to accept any other connections on the target host. With this update, the underlying code has been modified to delay the migration connection until the SPICE client is connected to the target destination. The guest virtual machines can now be successfully migrated without disrupting the display during the migration.
BZ#701106
Previously, migration of a virtual machine failed if the machine had an ISO image attached as a CD-ROM drive and the ISO domain was inactive. With this update, libvirt introduces the new startupPolicy attribute for removable devices, which allows marking CD-ROM and diskette drives as optional. With this option, virtual machines can now be started or migrated without a removable drive if the source image is inaccessible.
BZ#725373
When a destination host lost network connectivity while a domain was being migrated to it, the migration process could not be canceled. This update implements an internal keep-alive protocol, which is able to detect broken connections or blocked libvirt daemons. When such a situation is detected during migration, libvirt now automatically cancels the process.
BZ#729694
With certain combinations of IDE and VirtIO disks, a guest operating system did not boot after the installation process. This happened because the order of disks in which they were presented to the guest during the installation was different from the order used after the installation. As a result, the system could have been installed on a disk which was not used as the primary bootable disk. With this update, libvirt makes sure that the order in which disks are presented to a guest operating system during the installation is the correct order that will be used later once the guest operating system is installed.
BZ#729940
Previously, libvirt did not provide any way to prevent multiple clients from accessing a console device. When two clients connected to a single console of a guest, the connections entered a race condition on reading data from the console device. Each of the connections only got a fragment of the data and that fragment was not copied to the other connection. This rendered the terminal unusable to all the simultaneous connections. With this update, when opening a console, a check is performed to ensure that only one client is connected to it at a given point in time. If such a session is locked, a new connection has the ability to disconnect previous console sessions. Users are now able to safely access the console and disconnect inactive sessions to take control of a guest in case the console is accidentally left connected.
BZ#769503
Virtualization hosts can have thousands of CPUs and run a thousand guests, and libvirt should be capable of controlling all of them. However, libvirt was not able to do so, and the limit was below 1000, and users were therefore unable to fully utilize their hardware. With this update, the array of file descriptors which is passed to the child process is now allocated dynamically and can handle as many file descriptors as possible. Moreover, init and startup scripts have been changed so that the maximum limit of open files can be overridden for the libvirtd daemon. Users can now fully utilize their hardware and run as many guests as they require.
BZ#746666
Due to several problems with security labeling, libvirtd became unresponsive when destroying multiple guest domains with disks on an unreachable NFS storage device. This update fixes the security labeling problems and libvirtd no longer hangs under these circumstances.
BZ#795305
When live migration of a guest was terminated abruptly (using the Ctrl+C key combination), the libvirt daemon could have failed to accept any future migration request of that guest with the following error message:
error: Timed out during operation: cannot acquire state change lock
This update adds support for registering cleanup callbacks which are called for a domain when a connection is closed. The migration API is more robust to failures, and if a migration process is terminated, it can be restarted with a subsequent command.
BZ#752255
Previously, libvirt's implementation of nwfilter attempted to execute a temporary file generated directly in the /tmp/ directory, which failed if /tmp/ was mounted with the "noexec" options for security reasons. The implementation of nwfilter has been improved to avoid the need for a temporary file altogether, so it is no longer necessary for libvirt to modify or use files in the /tmp/ directory.
BZ#575160
Prior to this update, QEMU did not provide a notify mechanism when a block device tray status was changed. As a consequence, libvirt was unable to determine if the block data medium was ejected or was not present inside a guest. If the medium was ejected inside a guest, libvirt started the guest with the media being still present when migrating, saving and restoring the guest. This update introduces a new XML attribute for removable disk devices to represent and update the tray status.
BZ#758026
Under certain circumstances, a rare race condition between the poll() event handler and the dmidecode utility could occur. This race could result in dmidecode waiting indefinitely to perform a read operation on the already closed file descriptor. As a consequence, it was impossible to perform any tasks for virtualized guests using the libvirtd management daemon, or perform certain tasks using the virt-manager utility, such as creating a new virtual machine. This update modifies the underlying code so that the race condition no longer occurs and libvirtd and virt-manager work as expected.
BZ#758870
The libvirtd daemon could become unresponsive when starting the QEMU driver because the dmidecode tool needed a lot of time to process a large amount of data. It was consequently impossible to connect to the QEMU driver. The underlying source code has been modified to properly handle the POLLHUP event, so that users can now connect to the QEMU driver successfully in this scenario.
BZ#767333
The management application can request a guest to shut down or reboot. However, this was previously implemented by issuing Advanced Configuration and Power Interface (ACPI) events to a guest which could have ignored them. Consequently, the management application was unable to reboot such a guest. This update implements support for the guest-agent application that runs on a guest and calls the "shutdown" or "reboot" command when required. This means that a guest can be shut down or rebooted even when the guest ignores the ACPI events.
BZ#754128
When shutting down, a virtual machine changed its status from the "Up" state to the "Paused" state before it was shutdown. The "Paused" state represented the state when the guest had been already stopped, but QEMU was flushing its internal buffers and was waiting for libvirt to kill it. This state change confused users so this update adds respective events and modifies libvirt to use the "shutdown" state. A virtual machine now moves from "Up" to "Powering Down" and then to the "Down" state.
BZ#733587
If a domain failed to start, the host device for the domain was re-attached to the host regardless of whether the device was used by another domain. The underlying source code has been modified so that the device that is being used by another domain is not re-attached.
BZ#726174
Differences between the Red Hat Enterprise Linux and Debian implementations of the "nc" command, such as the presence or absence of the "-q" option, could lead to various problems. For example attempting to use a remote connection from a client expecting certain behavior to a server providing another behavior could fail on reconnection. With this update, libvirt probes capabilities of the "nc" command, and uses the appropriate options of the server even if the options differ from the "nc" on the client, which allows for successful interaction between either type of operating system.
BZ#771603
In Red Hat Enterprise Linux 6.2, libvirt unconditionally reserved PCI address 0:0:2.0 for a VGA adapter. Any domain that was created using an earlier version of libvirt with no VGA adapter and had another PCI device attached at this address could not be started. With this update, libvirt does not automatically use this PCI address for any device except for a VGA adapter. However, other devices can be attached at this address explicitly (either by the user or by using an older version of libvirt) and libvirt does not forbid to start domains with such devices. Thus, domains that could not be migrated from Red Hat Enterprise Linux 6.1 to 6.2 can be migrated from Red Hat Enterprise Linux 6.1 to 6.3.
BZ#782457
Previously, QEMU only offered the ability to perform a live snapshot of one disk at a time, but with no rollback functionality if the snapshot process failed. With this update, libvirt has been enhanced to take advantage of QEMU improvements that guarantee that either all disks have a successful snapshot, or that the failure is detected before any change which cannot be rolled back is made. This is easier for management applications performing a live disk snapshot of a guest with multiple disks.
BZ#697808
Parsing an XML file containing an incorrect root element caused an incorrect and confusing error to be displayed. The error message has been modified to display proper and detailed information about the problem when the user provides an incorrect XML file.
BZ#815206
If the umask used when starting init services was set to mask the executable or the search bit for other users, KVM virtual machines that were explicitly configured to use the "hugepages" mechanism could fail to start because the QEMU user was unable to access the directory that libvirt had created for QEMU in the hugetlbfs file system. This was because while the directory itself was owned by QEMU, its parent directory was not searchable by QEMU. To prevent this problem, when creating the parent directory, libvirt now makes sure that the parent directory is searchable by anyone regardless of umask settings.
BZ#796526
Previously, libvirt returned guest memory values in kibibytes (multiples of 1024), but with no indication of the scale. Furthermore, the libvirt documentation referred to kilobytes (multiples of 1000). Also, QEMU used mebibytes (multiples of 1024*1024) and these differences in scale could result in users making mistakes, such as giving a guest 1000 times less memory than planned, with a failure mode that was not easy to diagnose. Now the output is clear on the unit used, and the input allows users to use other units that can be more convenient.
BZ#619846
Previously, the qemu monitor command "query-migrate" did not return any error message when a problem occurred. Consequently, libvirt produced the "Migration unexpectedly failed" error message, which did not provide the proper information about the problem. The "fd:" protocol is now used to retrieve and produce the exact error message when a problem occurs.
BZ#624447
In some configurations, log messages similar to the following could be reported to libvirt or Red Hat Enterprise Virtualization users:
warning : virDomainDiskDefForeachPath:7654 : Ignoring open failure on xxx.xxx
These messages were harmless and could be safely ignored. With this update, the messages are no longer reported unless a problem occurs.
BZ#638633
Previously, libvirt and virsh ignored any script file given in the specification for a network interface of a type that did not actually use script files. To avoid confusion, this is now explicitly prohibited, an error is logged, and attempting to specify a script file for an interface type that does not support script files fails.
BZ#726771
This update provides improvements in reporting errors in XML file parsing, which makes identifying of errors easier.
BZ#746111
The libvirt package was missing a dependency on the avahi-libs package. The dependency is required due to libvirt linking in libavahi-client for mDNS support. As a consequence, the libvirtd daemon failed to start if the libvirt package was installed on the system without the avahi-libs package. With this update, the dependency on avahi-libs is now defined in the libvirt.spec file, and avahi-libs is installed along with libvirt.
BZ#802856
In previous versions of Red Hat Enterprise Linux, a "hostdev" device could be hot plugged to a guest, but making that device persistent across restarts of the guest required separately editing the guest configuration. This update adds support for persistent hot plug of "hostdev" devices, both to the libvirt API and to the virsh utility.
BZ#806633
Previously, attempting to migrate a server from a bridge network to direct network could fail when using libvirt with a virtio network interface. With this update, if a virtual guest created using the tools in Red Hat Enterprise Linux 6.2 or earlier is started on a host running Red Hat Enterprise Linux 6.3 with the vhost-net driver module loaded, and if that guest has a virtio network interface that uses macvtap, the "merge receive buffers" feature of the virtio driver is disabled. Compatibility with Red Hat Enterprise Linux 6.2 hosts is preserved and migration no longer fails under these circumstances.

Enhancements

BZ#761005
With this update, libvirt now supports for the latest Intel processors and new features these processors include.
BZ#767364
With this update, libvirt now supports family 15h microarchitecture AMD processors.
BZ#643373
Now, libvirt is capable of controlling the state (up or down) of a link of the guest virtual network interfaces. This allows users to perform testing and simulation as though plugging and unplugging the network cable from the interface. This feature also lets users isolate guests in case any issues arise.
BZ#691539
This update adds the ability to assign an SR-IOV (Single Root I/O Virtualization) network device Virtual Functions (VF) to a guest using the "interface" element rather than the "hostdev" element. This gives the user the opportunity to specify a known or fixed MAC address (<mac address='xx:xx:xx:xx:xx:xx'/>).
BZ#638506
Previously, the only way to perform storage migration was to stop a guest, edit the XML configuration file, and restart the guest. This led to a downtime that could have lasted several minutes. With this update, it is now possible to perform live storage migration with minimal guest downtime. This is ensured by new libvirt API flags to the virDomainStorageRebase() function, which map to new QEMU features.
BZ#693842
Previously, libvirt was able to notify a switch capable of the 802.1Qbg standard about changes in the guest network interface configuration, but there was no way for the switch to notify libvirt. This update provides extended support for libvirt synchronization with the lldpad daemon. As a result, if there are changes in the network infrastructure that require libvirt to re-associate the guest's interface, libvirt is informed and can take the proper action.
BZ#782034
With this update, libvirt supports a new model for the Small Computer System Interface (SCSI) controller, virtio-scsi.
BZ#713170
With this update, "fabric_name" of the "fc_host" class is exposed, so that users can see which fabric the virtual host bus adapter (vHBA) is connected to.
BZ#715019
This update introduces a new API, which allows the management system to query the disk latency using libvirt.
BZ#725013
It is sometimes required not only to delete a domain's storage but also overwrite the data to make sure sensitive data are no longer readable. This update introduces a new API, that allows users to erase the storage and use various wiping patterns.
BZ#769930
With this update, libvirt supports dynamic NUMA tuning, so that significant processes can be pre-bound to nodes with sufficient available resources.
BZ#740375
Previously, when doing disk snapshots, the guest had to be paused in order to avoid writing of data. Otherwise, the data could be corrupted. A new utility, guest-agent, has been introduced, and allows to freeze disks or file systems from inside the guest. It is no longer needed to pause the guest. However, disk write operations are delayed until the snapshot is completed.
BZ#768450
Previously, no mappings were specified for the "cpu64-rhel*" CPU models found in QEMU and therefore they could not be used. These mappings have been added with this update.
BZ#754073
Previously, it was not possible to see the memory used by the qemu-kvm process using only the virsh utility. The API call that reports the domain memory statistics has been modified to show this value. The value is now displayed when running the "virsh dommemstat" command.
BZ#533138
This update adds support for hot plugging and unplugging processors. It is now possible to add CPUs to guests and remove them as needed, without shutting down the guest.
BZ#713932
This update introduces a new virsh command, "change-media", which makes it easier to frequently insert and eject media from CD-ROM or floppy devices.
BZ#720691
The libvirt-guests init script attempted to make calls to the libvirtd daemon even if the daemon was inaccessible. As a consequence, the init script printed superfluous error messages that could be confusing. With this update, the script checks for a working connection, and skips calls on that connection if it is not working.
BZ#714759
This update introduces a new virsh command, "domiflist" to display detailed network interfaces information, and two new field for the "domblklist" command.
BZ#781562
Along with the "rombar" option that controls whether or not a boot ROM is made visible to the guest, QEMU also has the "romfile" option that allows specifying a binary file to present as the ROM BIOS of any emulated or pass-through PCI device. This update adds support for specifying "romfile" to both pass-through PCI devices, and emulated network devices that attach to the guest's PCI bus.
BZ#681033
Previously, libvirt did not provide means to add and display host metadata while listing guests. It was therefore impossible to store additional information about guests. A new element has been added to the libvirt XML configuration file, which allows users to store a description along with the API that allows modifications of guest metadata. The "virsh list" command has been updated to allow printing of the short description. As a result, identification of guests is now easier.
BZ#605953
This update adds a new virsh command, "iface-virsh", that allows users to "bridge" one of the host's Ethernet devices so that virtual guests can be connected directly to the physical network, rather than through a libvirt virtual network. The "iface-unbridge" command can be used to revert the interface to its previous state.
All users of libvirt are advised to upgrade to these updated packages, which fix these issues and add these enhancements.