RHEA-2011:1782 - Product Enhancement Advisory
new packages: vdsm
Product Enhancement Advisory
New vdsm packages are now available for Red Hat Enterprise Linux 6.
The Virtual Desktop Server Manager (VDSM) is a management module that serves as
a Red Hat Enterprise Virtualization Manager agent on Red Hat Enterprise
Virtualization Hypervisor or Red Hat Enterprise Linux 6 hosts.
Important: VDSM now controls qemu-kvm processes using libvirt. This is the most
significant change in these updated packages.
Additionally, VDSM now configures libvirt so that connections to its local
read-write UNIX domain socket are password-protected using Simple Authentication
and Security Layer (SASL). All operations that can change the state of virtual
machines (VM) on a Red Hat Enterprise Virtualization-controlled host must now be
performed from the Red Hat Enterprise Virtualization Manager.
This enhancement update adds the cmake package to Red Hat Enterprise Linux 5.
This package fixes the following bugs:
- In previous versions, the /root/.ssh/authorized_file keys on the host could be
erased when adding a Red Hat Enterprise Linux host to Red Hat Enterprise
Virtualization Manager. This version adds the keys for Red Hat Enterprise
Virtualization Manager to the list of authorized keys without erasing the list
- In previous versions, Windows guests with virtio disk driver from Red Hat
Enterprise Virtualization 2.2 became unresponsive in the Blue Screen of Death on
boot. As a workaround, the user can now start up the guest with the custom
property vioDiskCache=writethrough, install a newer virtio driver, and run the
VM again with the vioDiskCache custom property dropped. (BZ#706922)
- In previous versions, the gateway was not reported to Red Hat Enterprise
Virtualization Manager and left an empty field in the user interface when a VDSM
host (both Red Hat Enterprise Virtualization Hypervisor and VDSM on Red Hat
Enterprise Linux 6) was configured to acquire the IP address using Dynamic Host
Configuration Protocol (DHCP). With this update, VDSM reports the gateway in all
cases. Now, the GUI looks consistent. (BZ#716692)
- In previous versions, attempting to create or extend a storage domain on or
with a device that exposes a block size that is not exactly 512 bytes resulted
in a non-specific error message, which could confuse the user. With this
version, an error message correctly indicates that the device block size is not
supported. Now, the user gets the right information and can change the devices.
- In previous versions, VDSM mounted NFS using soft mounts with a timeout of 60
seconds and two retries. As a consequence, the I/O operations failed permanently
if a transient error occurred that was longer than 2 minutes and the VM paused
until the user manually resumed and the I/O operations were retried. With this
version, VDSM mounts with a timeout of 600 and 6 retries. (BZ#737181)
- In previous versions, the /root/.ssh directory was created with the wrong
SELinux context if the directory was missing from a host. As a consequence,
logging into the host using the Secure Shell (SSH) resulted in a denial. With
this update, the VDSM installation script creates the directory with the correct
This package also adds the following enhancements:
- In previous versions, VMs with VDisks over block devices ran using threads for
the asynchronous Input-Output (aio) model. With this version, VDSM starts the
VMs with the native aio for block device backed vdisks to achieve better
performance in this scenario. (BZ#611206)
- With this version, VDSM no longer assumes that network interfaces are named
eth[n], where [n] is a sequential integer value. This makes VDSM compatible with
the new device naming convention. (BZ#682790)
- With this version, VDSM configures libvirtd to run as an Upstart job, instead
of a SysV service to ensure that libvirtd is restarted in the case that it
terminates unexpectedly due to a segmentation fault. (BZ#694026)
All users who require VDSM are advised to install these new packages.
Before applying this update, make sure all previously-released errata relevant
to your system have been applied.
This update is available via the Red Hat Network. Details on how to use the Red
Hat Network to apply this update are available at
- Red Hat Virtualization 3 for RHEL 6 x86_64
- BZ - 598533 - selinux denies qemu access to monitor and vmchannel sockets
- BZ - 622446 - [vdsm] [libvirt] vm 'split brain' when 'vdsm' loose connection to 'libvirtd' after migration
- BZ - 660598 - [vdsm] failed migration leaves vm in pause although qemu is already dead
- BZ - 660665 - [vdsm] migration create returns successfully even though prepare volume fails
- BZ - 661321 - [vdsm] [scale] vdsm CPU consumption goes between 180-400 when running 100 vms
- BZ - 664510 - sosreport vdsm plugin duplicates device mapper plugin already in sosreport
- BZ - 668781 - SuperVDSM takes more then 2 second to create the socket when the system is congested
- BZ - 670005 - NFS domains report old lastCheck time when one is stuck
- BZ - 670432 - [vdsm] [storage] migrate master scenario - vdsm use old values in case restart takes places during operation
- BZ - 672208 - [vdsm] When migrating multiple VMs some getting stuck in Paused status.
- BZ - 673806 - Add config value for default migration bandwidth
- BZ - 677879 - [VDSM]- vdsm does not report an error when editing interface's ip address to an already used address.
- BZ - 677901 - [vdsm] when addNetwork is initiated with invalid bonding options no error is propagated.
- BZ - 677974 - [vdsm][watchdog]When libvirt dies vdsm watchdog will stop trying to restart after 30 seconds
- BZ - 680034 - [vsdm][storage][error-handlnig]connectStorageServer takes about 10 minutes during storage disconnection on certain topologies
- BZ - 681457 - vdsClient: usage not printed for vdsClient upgradeStoragePool command
- BZ - 682513 - VDSM - Storage: VDSM should work with new mke2fs default of discard=enabled
- BZ - 682790 - Support new Ethernet naming standards (from biosdevname)
- BZ - 684584 - [RHEVM][VDSM]POOL_DESCRIPTION key was missing from metadata (KeyError: 'POOL_DESCRIPTION' ; Storage.StoragePool::(getInfo) Pool metadata error)
- BZ - 684764 - VDSM: Can't run VM. due to Could not load certificates from /etc/pki/libvirt-spice/server-cert.pem
- BZ - 687403 - VDSM Error during deactivateStorageDomain while diconnecting network between host and ISO domain
- BZ - 688616 - VDSM should disable cpu cgroups
- BZ - 690079 - [vdsm][storage]ConnectStoragePool (HSM) fails on a large scale deployment
- BZ - 691340 - Change the storage version number from 23 to 2
- BZ - 692089 - [vdsm][storage]Vdsm doesn't clean-up stale lvs after copyImage operation fails
- BZ - 692149 - [vdsm] Volume operations can deadlock themselves
- BZ - 693157 - [vdsm] vdsm hangs forever after - prepareForShutdown returns 'OK'
- BZ - 693209 - getVGList fails when there are no VGs
- BZ - 693283 - [vdsm] [Snapshot] Merge snapshots fails on nfs storage domain
- BZ - 693424 - [vdsm][bootstrap] kvm-tools package should be installed during the bootstrap.
- BZ - 693991 - vdsm is missing support for multiple spice monitors
- BZ - 694026 - [vdsm][libvirt] as part of Libvirtd's upstart job- vdsm should initctl Libvirtd.
- BZ - 695057 - [vdsm][scale] when ~150 vms are running on a host and restarting libvirtd, prepareForShutdown takes ~35 minutes.
- BZ - 695244 - [vdsm] TypeError: 'NoneType' object is unsubscriptable as block Stats returns null
- BZ - 695355 - [vdsm][Storage] repoStats reports inaccessible storage domain as valid.
- BZ - 695393 - VDSM: Cannot stop VM that was paused due to storage I/O errors as long as the storage remains unavailable
- BZ - 698218 - [vdsm] processPool processes does not always exit when vdsm dies
- BZ - 698280 - Bad permissions on image after VM is imported by vdsm22
- BZ - 699976 - blocked on qemu-img: SPM fences itself during creation of several templates
- BZ - 700004 - [vdsm][Storage] vgtag metadata double checksum.
- BZ - 700048 - VDSM (w/ libvirt): when sending setVmTicket right after VM destroy libvirt returns empty reply which results in vdsm error
- BZ - 700442 - There are createVolume/copyImage tasks that stuck in "running" status
- BZ - 701107 - VDSM: guest with CD attached on boot will cause 'Stats function failed' error if disabling ISO during migration
- BZ - 701132 - [vdsm] SPM start fails as host fails to acquire resource although resource it not locked
- BZ - 701245 - VDSM - SpmStart command fails for the first time when switching SPM
- BZ - 701671 - Removing the iSCSI storage domain with host other that created it makes it not-attachable...
- BZ - 702275 - [vdsm] vdsm doesn't terminate on SIGTERM / sigusr1
- BZ - 702400 - [vdsm] getStoragePoolInfo takes more than 3 minutes to return in case non-master domain is blocked in NFS
- BZ - 702906 - [vdsm] VDSM finish its process pool on several NFS scenarios (start several VMs at once\block ISO domain)
- BZ - 703095 - Failed to start VM on rhevh since vdsm stopped cgroup without restarting libvirtd.
- BZ - 703848 - VDSM - Migrate VM fails when displayNetwork is used
- BZ - 704182 - [VDSM] [Storage] - If storage domain attach fails to update domain metadata operation is not recovered correctly
- BZ - 704666 - Thin provisioned VM disk on iSCSI gets inflated after migrating
- BZ - 704829 - VDSM:Can't start spm after vdsm lost it
- BZ - 705282 - VDSM: VdsClient create acpiEnable default value is wrong
- BZ - 705284 - VDSM: "Machine Destroyed" message when trying to set VM ticket
- BZ - 705322 - Can't start VM with spice console due to /etc/pki/vdsm/libvirt-spice/ rights
- BZ - 705839 - VDSM: Unexpected exception in backend when sending setVmTicket right after VM destroy
- BZ - 706085 - VDSM: After block connection from single host to ISO and tries to detach it the domain is left in an unteachable state
- BZ - 706817 - [vdsm] [Storage] Corrupted template from RAW preallocated disk.
- BZ - 706847 - [vdsm][Init-script] vdsmd's System v init script (stop) does not stop all vdsm processes.
- BZ - 706922 - VDSM - Add custom property for Disk-VirtIO-Cache
- BZ - 707217 - blocked on lvm scalability: vdsm fails to connect to storage pool for 5 minutes - on a loaded environment
- BZ - 707264 - [vdsm][storage]race condition on _recoverExistingVms
- BZ - 707539 - [vdsm] [init-script] vdsmd configures libvirt every time it starts up.
- BZ - 707587 - [vdsm][Networking] addNetwork with BOOTPROTO=dhcp fails
- BZ - 708022 - [vdsm]vdsm hsm sends lots of lvExtend though there is no need for it
- BZ - 708042 - New host can't be registered to RHEVM due to 'makeBridge' err - missing file 'addNetwork'
- BZ - 708773 - VDSM: if umount fails during create domain the storage is created but the user will get "can't find master domain" error
- BZ - 708805 - [vdsm] [Storage] File preallocated raw image dose not allocate the whole space.
- BZ - 709063 - [vdsm] [Storage] copyImage rollback fails with "OSError: Operation not permitted"
- BZ - 709102 - vdsm ignores pthread_mutex_init() errors
- BZ - 709382 - VDSM: Shared ISO Domain metadata corruption
- BZ - 709788 - [vdsm] [libvirt] 'vdsm' loose connection to 'libvirtd' and continues with migration
- BZ - 709998 - vdc_host_port should not be constraint to 443
- BZ - 710913 - [vdsm] in case virtIO & IDE are used, VDSM builds XML with both disks as 'hda'
- BZ - 713145 - [vdsm][storage]vdsm is writing duplicated pvs (different mapoffset) in the metadata
- BZ - 713185 - [vdsm] getStorageDomainsList displays the same domain twice.
- BZ - 713196 - VDSM - Storage: getSpmStatus doesn't return latest info
- BZ - 713215 - createStorageDomain fails on attribute error, if underlying devices already exist from a former same-structured domain.
- BZ - 713659 - [vdsm] When libvirt is non responsive and restarting libvirtd prepareForShutdown hangs forever.
- BZ - 713823 - few of concurrent createVolume calls fail when trying to hibernate multiple VM's
- BZ - 715327 - vdsm: cannot login to node when there are more them 190 targets in iscsi db
- BZ - 716675 - ValueError when formatting time-zone to one that does not exist in TZ environment variable.
- BZ - 716833 - Snapshot creation fails because image is not a leaf
- BZ - 717658 - Missing Storage Domains after connecting a new Srorage Server
- BZ - 717847 - There are virDomainGetBlockInfo errors when vdsmd restarts
- BZ - 717952 - [vdsm] moveImage with force Override to vm with 10 disks to NFS export domain always fails.
- BZ - 718993 - [vdsm][storage]refreshStoragePool takes a long time (more than default rhevm timeout) on large scale
- BZ - 719003 - [text] [vds_bootstrap] typo: "skippping" instead of "skipping"
- BZ - 719255 - [vdsm] when prepareVolume fails the image lock is not released.
- BZ - 719301 - [vdsm] [text] typo in vdsm logs: "aquired" instead of "acquired"
- BZ - 719346 - [RHEL.6][VDSM] - vdsClient command returns errors
- BZ - 719516 - [vdsm][vdsClient] exeption when executing list table just after or while stopping vms.
- BZ - 720257 - VDSM: When no connection between vdsm and storage domain teardownVolume failed
- BZ - 720355 - getStorageDomainsList fails if run before getDeviceList
- BZ - 720359 - [vdsm] [Init-script] vdsmd startup hangs forever while libvirt-guests service is running.
- BZ - 720981 - fail to create a SAN data domain(FC/iSCSI) in rhevm 3.0
- BZ - 721399 - [vdsm][performance]After fencing vdsm it takes it 4 minutes to come up
- BZ - 722649 - VDSM: CannotFormatAttachedStorageDomain error message when trying to remove data domain ( format=True )
- BZ - 722754 - [vdsm][error-handling][lvm-conf]vdsm should add disable_after_error_count in lvm.conf
- BZ - 723198 - No timeout on migration when connection between hosts crashes
- BZ - 723579 - [vdsm] PrepareForShutdown may hang forever when single vm is running and restarting libvirtd.
- BZ - 725240 - [vdsm][libvirt]libvirt logrotate should be done more frequently
- BZ - 725802 - vdsmd becomes defunct when blocking traffic between SPM to Master Storage Domain
- BZ - 725963 - [vdsm] isBlockDevice() throws an error if device doesn't exist` [getDeviceList fails]
- BZ - 725992 - Restart vdsm-reg immediately after configure RHEV-M on UI
- BZ - 726105 - [vdsm] No check for group permissions during creation of file data domain
- BZ - 726359 - use SIOCGIFFLAGS if operstate is unknown
- BZ - 726400 - ConnectStoragePool (HSM) with 50 FC SDs is taking lot of time
- BZ - 726410 - vdsm didn't filter out partitioned devices in getDeviceList.
- BZ - 726630 - [vdsm] Permission denied for metadata in mixed data-center
- BZ - 726941 - host installation failed - vds_bootstrap imports utils from yum
- BZ - 726949 - VDSM:Prepare volume fail (after connection to iSCSI storage is blocked and unblocked)
- BZ - 726952 - [Vdsm] race: trying to create snapshot after first creation failed because of vdsm restart will fail with error "There is no leaf in the image"
- BZ - 727537 - getDeviceList does not invalidate lvm cache before checking devices
- BZ - 727602 - Disable ticketing until libvirt solves bug 725009
- BZ - 727871 - Bad message parsing when reading from virto-serial socket
- BZ - 730297 - VM Lifecycle - VDSM must make sure that on setVmTicket currently connected user is disconnected
- BZ - 730912 - VDSM: Cannot remove all templates with 'Logical volume does not exist' error
- BZ - 730946 - Cannot discover iSCSI HBA luns in RHEVMstorage tab
- BZ - 731598 - No route to VDC error message during RHEL host install
- BZ - 732272 - Bonding fails - when nics order is not sorted
- BZ - 732275 - VDSM - Storage: Blocking NFS master domain causes other domains to be reported as having 0 free space
- BZ - 732652 - [vdsm][storage]validateStorageServerConnection fails in case mount is already exist on the host
- BZ - 732850 - BETA2 - Change defaults in RHEV-M registration dialog
- BZ - 732914 - vdsm][libvirtconnection] vdsm does not recover when it fails to connect to libvirtd upon startup.
- BZ - 733669 - report a specific migration on failure due to hostname lookup failure
- BZ - 733909 - [vdsm][storage]getSpmStatus returns invalid spmId
- BZ - 733910 - __cleanStorageRepository fails to clean links in directories
- BZ - 734768 - If kvm is not installed or loaded getVdsCaps fails
- BZ - 735932 - vds_bootstrap: do not download and do not use redhat-lsb
- BZ - 736103 - [vdsm][resourceFactories] when prepareVolume gets timeout error on acquire exclusive resource, it doesn't fail.
- BZ - 736154 - Add 'set rhevm admin password' to RHEVH config TUI and kernel cmdline args
- BZ - 736422 - VDSM - should be able to recover from malformed message from rhev-agent
- BZ - 736645 - VDSM - Storage: reconstructMaster on NFS takes longer than RHEVM's timeout
- BZ - 737329 - Connect storage pool with inaccessible master domain should raise StoragePoolMasterNotFound
- BZ - 737722 - vdsm-reg params are not removed from cmdline after registering to RHEV-M
- BZ - 741190 - VDSM: /etc/logrotate.d/libvirtd contains bad data, causes e-mail flood from cron
- BZ - 742267 - RHEV 3 can no longer use an NFS export which was usable by RHEV 2.2
- BZ - 742764 - Can't create VM from template after SPM failover.
- BZ - 743887 - Too many open files error makes hypervisor marked unresponsive
- BZ - 744549 - [vdsm][caps] vdsm fails to start if more than a single redhat-release package are installed
- BZ - 746975 - VDSM - Storage: Can't deactivate unreachable NFS domain
Red Hat Virtualization 3 for RHEL 6