qemu-server (8.0.10) bookworm; urgency=medium
* sdn: pass vmid and hostname to allow requesting a new mapping
-- Proxmox Support Team <support@proxmox.com> Wed, 22 Nov 2023 14:12:46 +0100
qemu-server (8.0.9) bookworm; urgency=medium
* add clipboard option to to vga config entry
* api: add clipboard variable to return at status/current
* recommend libpve-network-perl for SDN support
* initial support for dhcp ip allocation in dhcp-enabled SDN zones
-- Proxmox Support Team <support@proxmox.com> Tue, 21 Nov 2023 15:40:27 +0100
qemu-server (8.0.8) bookworm; urgency=medium
* fix #2816: restore: remove timeout when allocating disks
* start: increase maximal timeout if using PCI passthrough
* fix #4522: api: vncproxy: also set environment variable for the ticket
if the websocket option is not set
* backup, migrate: fix races with suspended VMs that can wake up
* cpu hotplug: cannot change feature online, so keep these as pending change
* nbd-stop: increase timeout to 25s
* start: add warning if a deprecated machine version is configured
-- Proxmox Support Team <support@proxmox.com> Sun, 12 Nov 2023 18:54:37 +0100
qemu-server (8.0.7) bookworm; urgency=medium
* fix #4822: vzdump: fix PBS encryption for guests without disks
* qmeventd: fix parsing of VMID in presence of legacy cgroup entries
* api: check access for already configured bridge when updating vNIC
* fix #4620: make 'ide1' and 'ide3' drive keys work for machine type q35
* cloudinit: fix two checks that were mistakenly restricted to root only,
one for setting the ciupgrade option and one for updating the cloudinit
drive
* migration: improve format hint when allocating live-migrated disk on the
target to make e.g. remote-migration with qcow2 and LVM-thin target work
* net: fix setting value for tx_queue_size
* fix #3963: allow backup of template VM with immutable TPM drive
-- Proxmox Support Team <support@proxmox.com> Mon, 21 Aug 2023 11:30:45 +0200
qemu-server (8.0.6) bookworm; urgency=medium
* cloudinit: restore previous default for package upgrades
* migration: only migrate disks used by the guest, not also those that are
owned by them (VMID in name) but not referenced in the config
* migration: fail when aliased volume are detected, as referencing the same
volume multiple times can lead to unexpected behavior in a migration.
* migration: fix issue with qcow2 cloudinit disk live migration
-- Proxmox Support Team <support@proxmox.com> Wed, 21 Jun 2023 13:03:01 +0200
qemu-server (8.0.5) bookworm; urgency=medium
* restore: extend permissions checks
* vm start: always reset any failed-state of the VM systemd scope to avoid
failing a re-start after, e.g., a OOM kill.
-- Proxmox Support Team <support@proxmox.com> Wed, 21 Jun 2023 09:17:41 +0200
qemu-server (8.0.4) bookworm; urgency=medium
* vCPU config: add new x86-64-v2, x86-64-v3 and x86-64-v4 models
* fix #4784: helpers: cope with native versions in manager version check
* enable cluster mapped USB devices for guests
* enable cluster mapped PCI devices for guests
-- Proxmox Support Team <support@proxmox.com> Mon, 19 Jun 2023 07:24:11 +0200
qemu-server (8.0.3) bookworm; urgency=medium
* qemu: fix permission check call
-- Proxmox Support Team <support@proxmox.com> Fri, 09 Jun 2023 12:20:40 +0200
qemu-server (8.0.2) bookworm; urgency=medium
* cfg2cmd: use actual backend names instead of removed tty and paraport
aliases
* cfg2cmd: replace deprecated no-acpi option with acpi=off machine flag
* cfg2cmd: replace deprecated no-hpet option with hpet=off machine flag
* schema: avoid using deprecated -no-hpet in example for 'args' property,
instead pass thate via new machine option
* allow setting ipconfigX with VM.Config.Cloudinit
* fix #3428: cloudinit: add parameter for upgrade on boot
* cloudinit: fix 'pending' api endpoint
* fast plug options: add migrate_downtime and migrate_speed for convenience
* fix #517: api: allow resizing qcow2 disk with snapshots
* fix #2315: api: have resize endpoint spawn a worker task
* cloudinit: pass through hostname via fqdn field
* qmeventd: extract vmid from cgroup file instead of cmdline
* config: implement method to calculate derived properties from a config
* api: check bridge access for create, update, clone & restore
* qm: remote migration: improve error when storage cannot be found
-- Proxmox Support Team <support@proxmox.com> Fri, 09 Jun 2023 10:26:19 +0200
qemu-server (8.0.1) bookworm; urgency=medium
* fix #4737: qmeventd: gracefully handle interrupted epoll_wait call
* handle and warn about VM network interfaces not attached to any bridges
* block resize: avoid passing zero size to QMP command
* qmrestore: improve description of bwlimit parameter
* api: switch agent api call to 'array' type
* tests: fix invoking migration tests with make
-- Proxmox Support Team <support@proxmox.com> Wed, 07 Jun 2023 13:50:09 +0200
qemu-server (8.0.0) bookworm; urgency=medium
* never enable 'smm' flag for the 'virt' machine type (doesn't exist)
* test: mock calls that can fail in a chroot environment
* rebuild for Debian Bookworm based releases
-- Proxmox Support Team <support@proxmox.com> Fri, 19 May 2023 15:07:45 +0200
qemu-server (7.4-3) bullseye; urgency=medium
* backup prepare: fix format detection for disks without storage ID
* backup prepare: improve error messages
-- Proxmox Support Team <support@proxmox.com> Mon, 27 Mar 2023 11:17:16 +0200
qemu-server (7.4-2) bullseye; urgency=medium
* avoid list context for volume_size_info calls as otherwise we
unnecessarily take a slower code path
-- Proxmox Support Team <support@proxmox.com> Tue, 21 Mar 2023 16:51:01 +0100
qemu-server (7.4-1) bullseye; urgency=medium
* fix #4553: nvidia vgpu: reuse smbios uuid for '-uuid' parameter
* pci: workaround nvidia driver issue on mdev cleanup
* memory: hotplug: sort by numerical ID rather than slot when unplugging
* memory: use the DIMM list info from QEMU for unplug
-- Proxmox Support Team <support@proxmox.com> Mon, 20 Mar 2023 17:24:45 +0100
qemu-server (7.3-4) bullseye; urgency=medium
* fix #4378: standardized error for missing OVMF files
* schema: memory: be precise that unit is binary prefix
* close #2792: allow online migration with replicated snapshots
* schema: OS type: note that the l26 type is also compatible with Linux 6.x
* hotplug: disk: mark the 'aio' (async IO) as non-hotpluggable to avoid
suggesting that it already changed
* fix #4525: clone disk: disallow mirror if it might cause problems with
io_uring using the same heuristics as for start-up
* start: make not being able to set polling interval for ballooning
non-critical
* swtpm: enable logging to `/run/qemu-server/$vmid-swtpm.log`
* fix #4140: vzdump: transform the previous hardcoded behavior of issuing a
fs-freeze and fs-thaw if QGA is enabled into an overrideable option named
'fs-freeze-on-backup'
* update network dev: MTU is not hot-pluggable, avoid suggesting so
* fix #4249: make image clone or conversion respect bandwidth limit
-- Proxmox Support Team <support@proxmox.com> Thu, 23 Feb 2023 17:12:42 +0100
qemu-server (7.3-3) bullseye; urgency=medium
* rollback: ignore auto-start request if VM is already running
* memory hot-plug: check correct value for maximal memory check
* fix #4435: device list: avoid error for undefined value
* fix #4358: ignore any suspended lock when destroying a VM
* migration: log error from query-migrate, if any, upon migration failure
* cd rom handling: return a clearer error when there is no CD-ROM drive
* migration: nbd export: switch away from deprecated QMP command
-- Proxmox Support Team <support@proxmox.com> Mon, 16 Jan 2023 13:52:30 +0100
qemu-server (7.3-2) bullseye; urgency=medium
* fix #4372: improve edge-case for config-loading on VM resume when
migrating
* ovmf cmd assembly: re-work and re-order arguments assembly
-- Proxmox Support Team <support@proxmox.com> Fri, 16 Dec 2022 12:54:53 +0100
qemu-server (7.3-1) bullseye; urgency=medium
* vm resume: improve loading just recently moved config on nocheck/migrate
handling
-- Proxmox Support Team <support@proxmox.com> Mon, 21 Nov 2022 13:43:59 +0100
qemu-server (7.2-12) bullseye; urgency=medium
* config: only save unique tags when updating them via the API
* api: create/update vm: fix clamping CPU units function calls
-- Proxmox Support Team <support@proxmox.com> Mon, 21 Nov 2022 08:36:06 +0100
qemu-server (7.2-11) bullseye; urgency=medium
* fdb: only manage FDB entries for Linux bridges, ignore OVS for now
-- Proxmox Support Team <support@proxmox.com> Sun, 20 Nov 2022 16:30:28 +0100
qemu-server (7.2-10) bullseye; urgency=medium
* fix #4321: properly check cloud-init drive permissions, require both
VM.Config.CDROM and VM.Config.Cloudinit, and not VM.Config.Disk, for being
able to add a cloud init drive in the first place.
* api: config update: enforce new tag permission system when setting or
removing tags from a guest
* parse config: do not validate informative values in cloud init section
* fix edge-cases on new cloudinit pending/active recording
* mtunnel: add API endpoints
* migrate: add foundation for remote (external cluster) migration, add
respective endpoints and qm `remote-migrate` CLI command
* memory hotplug: make max-memory dynamically calculated from the physicall
address bits the VM will use, that is the actual one from the config, if
set, the one from the host for CPU type host and 40 bits as fallback for
everything else. Calculate the addressable memory (e.g., 40 bits = 1 TiB)
and half that for the possible max-memory a VM can use, using the previous
hard-coded 4 TiB as overall maximum for backward compat.
Admins with inhomogeneous CPUs and thus possible different bit-widths need
to take special care themselves to ensure that a VM with memory hot-plug
configured can run on other nodes, for example for live-migration.
-- Proxmox Support Team <support@proxmox.com> Thu, 17 Nov 2022 17:48:03 +0100
qemu-server (7.2-8) bullseye; urgency=medium
* fix #4296: virtio-net: enable packed queues for qemu 7.1
* virtio-net: increase defaults rx- and tx-queue-size to 1024
* fix #4296: virtio-net: enable packed queues for QEMU machines using 7.1 or
newer
* net: increase max queues to 64
* fix #4284: add read-only to non-hotpluggable disk options
* delay cloudinit generation in hotplug
* record cloud-init changes in the cloudinit section
* rework cloudint config pending handling
-- Proxmox Support Team <support@proxmox.com> Wed, 16 Nov 2022 18:23:39 +0100
qemu-server (7.2-7) bullseye; urgency=medium
* api: create/update vm: automatically clamp cpuunit value depending of
cgroup version
* improve cloud init support and add cloudinit hotplug
* vzdump: skip `special:cloudinit` section
* fix #3890 - GUI: warn for unlikely iothread config clashes
* fix #4228: add `start` parameter to snapshot rollback API so that one can
automatocally start the VM after rollback finished.
* vm start/stop: cleanup passed-through pci devices in more situations
* fix #3593: allow one to configure task set affinity for VMs
* fix #4324: USB: use qemu-xhci for machine versions >= 7.1
* usb: increase max USB devices from 5 to 14 for modern 7.1 machine
and OS versions (Linux 2.6+ annd Windows 8+)
* fix #4201: delete cloud-init disk on rollback
* net devs: register vNIC MAC-Address manually to FDB on start/resume if
bridge has learning disabled
-- Proxmox Support Team <support@proxmox.com> Sun, 13 Nov 2022 15:46:18 +0100
qemu-server (7.2-6) bullseye; urgency=medium
* schema: move 'pve-targetstorage' to pve-common
-- Proxmox Support Team <support@proxmox.com> Mon, 07 Nov 2022 16:22:50 +0100
qemu-server (7.2-5) bullseye; urgency=medium
* qmp client: increase guest fstrim timeout to 10 minutes
* fix #3577: prevent suspension for VMs with pci passthrough
* cpu config: map deprecated IceLake-Client CPU type to IceLake-Server
* snapshot: save VM state: propagate error from QEMU
* api: create disks: avoid adding secondary cloud-init drives
* vzdump: TPM state: escape drive string
* qmp client: increase default fallback timeout to 5s
* fix regex matching network devices in qm cleanup so that vNICs with double
digit IDs are covered too
* qmeventd: rework 'forced_cleanup' handling and set timeout to 60s
* qmeventd: send QMP 'quit' command instead of SIGTERM
* vzdump: set max-workers QMP option when specified and supported
* fix #4099: disable io_uring for virtual disks on CIFS storages for now
* qm: move VM-disk related commands to own command group, keep old ones
around for backward compatibility
-- Proxmox Support Team <support@proxmox.com> Mon, 07 Nov 2022 16:15:16 +0100
qemu-server (7.2-4) bullseye; urgency=medium
* fix #3754: encode JSON as utf8 for CLI
* cpuconfig: add amd epyc milan model
* fix #4115: enable option to name QEMU threads after their main purpose
* fix #4119: give namespace parameter to live-restore
* automatically add 'uuid' parameter when passing through NVIDIA vGPU
* vzdump/pbs: die with missing, but configured encryption key
* vzdump/pbs: die with missing, but configured master key
-- Proxmox Support Team <support@proxmox.com> Tue, 16 Aug 2022 13:59:20 +0200
qemu-server (7.2-3) bullseye; urgency=medium
* support pbs namespaces
-- Proxmox Support Team <support@proxmox.com> Thu, 12 May 2022 15:14:39 +0200
qemu-server (7.2-2) bullseye; urgency=medium
* api: reassign disk: drop moved disk from boot order
* explicitly check some prerequisites for virtio-gl display
-- Proxmox Support Team <support@proxmox.com> Mon, 02 May 2022 17:26:16 +0200
qemu-server (7.2-1) bullseye; urgency=medium
* migrate: add log for guest fstrim and make a failure noticable
* migrate: resume initially running VM when failing after convergence
* parse vm config: remove "\s*" from multi-line comment regex
* memory: enable balloon free-page-reporting for auto-memory reclaim
* enable spice also for virtio-gl and virtio-gpu displays and report so in
status API
* api: create: allow overriding non-disk options during restore
* fix #3861: migrate: fix live migration when cloud-init changes storage
-- Proxmox Support Team <support@proxmox.com> Thu, 28 Apr 2022 18:35:22 +0200
qemu-server (7.1-5) bullseye; urgency=medium
* avoid writing the config if there are no pending changes to apply
* fix #3792: cloudinit: use of uninitialized value
* pci: allow override of PCI vendor/device ids
* drive mirror monitor: warn when suspend/resume/freeze/thaw calls fail
* update config: allow setting boot-order and dev in one go
* migrate: move tunnel-helpers to pve-guest-common
* fix #3683: agent file-write: enable user to encode the content themselves
* cpu units: lower minimum for accessing full cgroupv2 range
* fix #3845: also clamp cpu units to cgroup dependent valid range on hotplug
* clone disk: force raw format for TPM state
* fix #3886: QEMU restore: verify storage allows images before writing
* fix #3733: bump the timeout used to wait that a for backup started VM is
fully stopped (i.e., it's "$vmid.scope vanished) to 20 seconds after the
backup has finished to
* qmp client: increase timeout for thaw to better accommodate the QGA running
in Windows VMs
* api: vm start: 'force-cpu' is for internal migration use only, mark as
such
* device unplug: verify that unplugging SCSI disk completed before
continuing with remaining unplug work.
* clone disk: remove ancient check for min QEMU version 2.7
* clone disk: pass in efi vars size rather than config
* clone disk: allow cloning from an unused or unreferenced disk
* parse ovf: untaint path when getting the file's size info
* image convert: allow block device as source
* fix #3424: api: snapshot delete: wait for active replication
* PCI: allow longer pci domains
* fix #3957: spell 'occurred' correctly
* clone disk: also clone EFI disk from snapshot
* api: add endpoint for parsing .ovf files
* api: support VM disk import
* migrate: keep VM paused after migration if it was before
* vga: add virtio-gl display type for VIRGL
* restore: cleanup oldconf: also clean up snapshots from kept volumes
* restore: also deactivate/destroy cloud-init disk upon error
-- Proxmox Support Team <support@proxmox.com> Mon, 25 Apr 2022 20:15:59 +0200
qemu-server (7.1-4) bullseye; urgency=medium
* migrate: send updated TPM state volume ID to target node on local-storage
migration
-- Proxmox Support Team <support@proxmox.com> Mon, 22 Nov 2021 17:07:13 +0100
qemu-server (7.1-3) bullseye; urgency=medium
* replication: do not setup NBD server on VM migrate for the TPM state,
QEMU cannot access it directly and we already migrate it via the non-QEMU
storage migration anyway.
-- Proxmox Support Team <support@proxmox.com> Tue, 16 Nov 2021 14:04:45 +0100
qemu-server (7.1-2) bullseye; urgency=medium
* cfg2cmd: disable SMM when display=none and SeaBIOS is both used
* pci: do not reserve pci-ids for mediated devices, already handled by sysfs
anyway
* exclude efidisk and tpmstate for boot disk selection heuristic
-- Proxmox Support Team <support@proxmox.com> Mon, 15 Nov 2021 16:59:23 +0100
qemu-server (7.0-19) bullseye; urgency=medium
* rollback: improve interaction with snapshot replication
* cli: qm: rename 'move_disk' command to 'move-disk' with an alias for
backward compatibility
* pi: move-disk: add possibility to reassign a disk to another VM
* turn SMM off when SeaBIOS and a serial-display are used in combination to
avoid a possible boot loop
-- Proxmox Support Team <support@proxmox.com> Thu, 11 Nov 2021 12:49:10 +0100
qemu-server (7.0-18) bullseye; urgency=medium
* use non SMM ovmf code file for i440fx machines
* fix hot-unplugging (removing) a cpulimit on a running VM
* vm start: only print tpm-related message if there is an actual instance
* vzdump: increase timeout for QMP 'cont' command after backup started
* drives: expose readonly flag for SCSI/VirtIO drives as 'ro' property
* qemu-agent: allow hotplug of the 'fstrim cloned disk' option
* fix #2429: allow to specify cloud-init vendor snippet via 'cicustom'
* config: add new meta property with the VM creation time
* config: meta: also save the QEMU version installed during creation
* cfg2cmd: switch off ACPI hotplug on bridges for q35 VMs with linux as
ostype to avoid changes in network interface naming due to systemd's
predicatble naming scheme
-- Proxmox Support Team <support@proxmox.com> Thu, 04 Nov 2021 15:29:55 +0100
qemu-server (7.0-17) bullseye; urgency=medium
* fix #3258: block vm start when a PCI(e) device is already in use
* snapshot: fix TPM state with RBD
* swtpm: wait for PID file to appear before continuing with VM start
* OS type: add entry for Windows 11/Server 2022
-- Proxmox Support Team <support@proxmox.com> Thu, 21 Oct 2021 11:57:09 +0200
qemu-server (7.0-16) bullseye; urgency=medium
* ovmf: support secure boot enabled code images
* ovmf: support provisioning an EFI vars template with secureboot by default
on and distribution + Microsofts secure-boot key pre-enrolled
-- Proxmox Support Team <support@proxmox.com> Tue, 05 Oct 2021 20:22:18 +0200
qemu-server (7.0-15) bullseye; urgency=medium
* api: return task-worker UPID in create template endpoint
* api: destroy VM: remove pending volumes as well
* fix #3075: add TPM v1.2 and v2.0 support via swtpm~
-- Proxmox Support Team <support@proxmox.com> Tue, 05 Oct 2021 07:24:52 +0200
qemu-server (7.0-14) bullseye; urgency=medium
* fix #3581: pass size via argument for memory-backend-ram QMP call
* fix #3608: improve removal of the underlying SCSI controller when removing
last drive on it
* migrate: do not suggest that we map shared storages to avoid that
subsequent checks could result in false negatives.
-- Proxmox Support Team <support@proxmox.com> Wed, 22 Sep 2021 09:31:06 +0200
qemu-server (7.0-13) bullseye; urgency=medium
* fix bootorder regression with implicit default order
-- Proxmox Support Team <support@proxmox.com> Thu, 5 Aug 2021 14:03:14 +0200
qemu-server (7.0-12) bullseye; urgency=medium
* fix #3371: import ovf: allow the use of dots in the VM name
* bootorder: fix double entry on cdrom edit
-- Proxmox Support Team <support@proxmox.com> Fri, 30 Jul 2021 16:53:44 +0200
qemu-server (7.0-11) bullseye; urgency=medium
* nic: support the intel e1000e model
* lvm: avoid the use of io_uring for now
* live-restore: fail early if target storage doesn't exist
* api: always add new CD drives to bootorder
* fix #2563: allow live migration with local cloud-init disk
-- Proxmox Support Team <support@proxmox.com> Fri, 23 Jul 2021 11:08:48 +0200
qemu-server (7.0-10) bullseye; urgency=medium
* avoid using io_uring for drives backed by LVM and configured for write-back
or write-through cache
-- Proxmox Support Team <support@proxmox.com> Wed, 07 Jul 2021 14:56:34 +0200
qemu-server (7.0-9) bullseye; urgency=medium
* cpu weight: always clamp value to lower maximum for cgroup v2 and fix
defaults (v1 -> 1024, v2 -> 100)
* api: improve error handling when applying pending config changes
-- Proxmox Support Team <support@proxmox.com> Wed, 07 Jul 2021 12:02:13 +0200
qemu-server (7.0-7) bullseye; urgency=medium
* improve #3329: ensure write-back is used over write-around for EFI disk,
as OVMF profits a lot from cached writes due to its frequent
read-modify-write operations
-- Proxmox Support Team <support@proxmox.com> Mon, 05 Jul 2021 20:49:50 +0200
qemu-server (7.0-6) bullseye; urgency=medium
* live-restore: preload efidisk before starting VM
* For now do not use io_uring for drives backed by Ceph RBD, with KRBD and
write-back or write-through cache enabled, as in that case some polling/IO
may hang in QEMU 6.0.
-- Proxmox Support Team <support@proxmox.com> Fri, 02 Jul 2021 09:45:06 +0200
qemu-server (7.0-5) bullseye; urgency=medium
* don't default to O_DIRECT (cache=none) on btrfs without nocow
* fix #2175: api: update VM: check old drive-config for permissions too to
ensure a valid transition when limited to CDROM changes.
-- Proxmox Support Team <support@proxmox.com> Thu, 24 Jun 2021 18:58:19 +0200
qemu-server (7.0-4) bullseye; urgency=medium
* enable io-uring support by default when running QEMU 6.0 or newer
* VM start: always check if storages of volumes support correct content-type
* use KillMode 'process' for systemd scope to cope with depreacation of
KillMode=none
* cli, api: handle new warnings task status
* improve backup of templates with EFI disks and with SATA and IDE
disk controllers in use
-- Proxmox Support Team <support@proxmox.com> Wed, 23 Jun 2021 12:57:27 +0200
qemu-server (7.0-3) bullseye; urgency=medium
* vzdump: add master key support
* vzdump: drop legacy fallback logging for dirty-bitmap
* vm destroy: do not remove unreferenced disks by default
* fix #3329: turn on cache=writeback for efidisks on rbd
* avoid setting LUN number for drives when the `pvscsi` controller is used,
as that cannot handle multiple LUNs, increase the `scsi-id` instead
* config: limit description/comment length to 8 KiB
* migrate: enforce that image content type is available and configured on
target storage
-- Proxmox Support Team <support@proxmox.com> Mon, 21 Jun 2021 11:17:52 +0200
qemu-server (7.0-2) bullseye; urgency=medium
* api: clone: sort vm disks to keep numbers consistent
* api: VM status: make template property optional in return object
* add compatibility for QEMU 6.0
* destroy VM: always remove (referenced) VM state volumes
* destroy VM: also check if unused volumes are base images
* live-restore: log more similar to regular restore, outputting the user the
PBS repo/snapshot and target for each drive.
-- Proxmox Support Team <support@proxmox.com> Fri, 28 May 2021 12:46:36 +0200
qemu-server (7.0-1) pve; urgency=medium
* re-build for Proxmox VE 7 / Debian Bullseye
-- Proxmox Support Team <support@proxmox.com> Thu, 13 May 2021 19:11:18 +0200
qemu-server (6.4-2) pve; urgency=medium
* fix #2862: allow sata/ide template backups
* migration: improve speed-limits for >1G connections again
* fix getting bootdisk size for new bootorder config scheme
-- Proxmox Support Team <support@proxmox.com> Thu, 29 Apr 2021 16:16:04 +0200
qemu-server (6.4-1) pve; urgency=medium
* fix the +pveN versioned machine types when PXE is used
* migration: avoid re-scanning all volumes
* migration: do not always set default speed limit if none is configured
* migration: rework logging to more humand friendly format, avoiding to much
output noise
* qmrestore: add live-restore option for CLI tool
* live-restore: hold 'create' lock during operation
* live-restore: don't remove VM on error, to allow an VM user to save any new
data before retrying the operation.
* fix #3369: auto-start vm after failed stopmode backup
-- Proxmox Support Team <support@proxmox.com> Fri, 23 Apr 2021 16:26:54 +0200
qemu-server (6.3-11) pve; urgency=medium
* enable live-restore tech preview for Proxmox Backup Server hosted backup
snapshots.
* drive mirror: rework periodic status reporting to be human friendlier
* drive mirror: stop logging progress for a disk once it got ready
* image convert: use human-readable units in progress report
-- Proxmox Support Team <support@proxmox.com> Thu, 15 Apr 2021 18:32:06 +0200
qemu-server (6.3-10) pve; urgency=medium
* increase timeout for block (disk) resize QMP command
* fix #3314: cloudinit: IPv6 requires type 'static6'
* fix #2670: cloudinit: enable SLAAC again now that client support is there
-- Proxmox Support Team <support@proxmox.com> Tue, 30 Mar 2021 18:40:58 +0200
qemu-server (6.3-9) pve; urgency=medium
* restore vma: fix applying storage-specific bandwidth limit
* snapshot: set migration caps before savevm-start
* vzdump: improve error logging for query-proxmox-support to avoid
false-positives
-- Proxmox Support Team <support@proxmox.com> Fri, 26 Mar 2021 09:47:27 +0100
qemu-server (6.3-8) pve; urgency=medium
* qm status: sort hash keys on verbose output
* improve windows VM version pinning on VM creation
-- Proxmox Support Team <support@proxmox.com> Fri, 12 Mar 2021 10:01:09 +0100
qemu-server (6.3-7) pve; urgency=medium
* vzdump: increase Proxmox Backup Server backup QMP command timeout
-- Proxmox Support Team <support@proxmox.com> Tue, 09 Mar 2021 08:21:43 +0100
qemu-server (6.3-6) pve; urgency=medium
* fix #3324: clone disk: use larger blocksize for EFI disk
* fix #3301: status: add currently running machine and QEMU version to full
status
* api: add endpoint to list all available QEMU machine type and version
tuples
* always pin virtual machines with Windows as ostype to a fixed QEMU machine
version by default. For existing VMs with Windows based OS-type use the 5.1
machine version (or the next available one, for older QEMU versions) to
improve stabillity of the hardware layout from Windows point of view. Linux
and other OS types are not as sensitive to those changes, so keep the
default to the currently latest available machine versions for non-Windows
VMs.
* update VM: check for CDROM not just drive permissions when removing a
device
-- Proxmox Support Team <support@proxmox.com> Fri, 05 Mar 2021 21:42:59 +0100
qemu-server (6.3-5) pve; urgency=medium
* cloudinit: add opennebula config format
* cloudinit: remove pending delete on online regenerate image
* snapshot/save-vm: periodically print progress and show information about
drives during snapshot
* qmeventd: explicitly close() pidfds
-- Proxmox Support Team <support@proxmox.com> Thu, 11 Feb 2021 18:05:18 +0100
qemu-server (6.3-4) pve; urgency=medium
* audio: add the "none" dummy audio backend
* fix drive-mirror completion with cloudinit
* vm destroy: allow opt-out of purging unreferenced disks
* fix #2788: do not resume vms after backup if they were paused before
* anchor CPU flag regex to avoid arbitrary flag suffixes
-- Proxmox Support Team <support@proxmox.com> Thu, 28 Jan 2021 17:21:07 +0100
qemu-server (6.3-3) pve; urgency=medium
* api: adapt VM destroy and purge description
* clone disk: fix regression in offline clone of efidisk
* cloudinit: fix cloning/restoring of cloudinit disks in raw format
-- Proxmox Support Team <support@proxmox.com> Tue, 15 Dec 2020 16:33:01 +0100
qemu-server (6.3-2) pve; urgency=medium
* PBS: use improved method to assemble repository url, fixing issues when
using IPv6 or non-default ports
-- Proxmox Support Team <support@proxmox.com> Thu, 03 Dec 2020 18:06:25 +0100
qemu-server (6.3-1) pve; urgency=medium
* deactivate volumes after storage migrate
* print query-proxmox-support result in 'full' status
* clone disk: avoid errors after disk was moved by QEMU
* replace cgroups_write by cgroup change_cpu_shares && change_cpu_quota
-- Proxmox Support Team <support@proxmox.com> Wed, 25 Nov 2020 14:30:50 +0100
qemu-server (6.2-20) pve; urgency=medium
* don't migrate replicated VM whose replication job is marked for
removal
* ensure qmeventd service is stopped after pve-guests and pve-ha-lrm service
on shutdown
-- Proxmox Support Team <support@proxmox.com> Thu, 12 Nov 2020 17:08:45 +0100
qemu-server (6.2-19) pve; urgency=medium
* fix #3113: unbreak drive hotplug
* qmeventd: add handling for -no-shutdown QEMU instances, to avoid errors if
the guest OS shuts down the VM during a backup job.
-- Proxmox Support Team <support@proxmox.com> Thu, 05 Nov 2020 13:37:00 +0100
qemu-server (6.2-18) pve; urgency=medium
* migrate: tell QEMU to enable dirty-bitmap migration, if supported
* partially fix #3056: always try to cancel backups when failed to start job
-- Proxmox Support Team <support@proxmox.com> Thu, 29 Oct 2020 18:23:13 +0100
qemu-server (6.2-17) pve; urgency=medium
* bootorder: don't print empty 'order=' property
-- Proxmox Support Team <support@proxmox.com> Thu, 22 Oct 2020 16:08:57 +0200
qemu-server (6.2-16) pve; urgency=medium
* fix #3010: add 'bootorder' parameter for better control of boot devices
* fix VM clone from snapshot with cloudinit disk
* fix various possible issues by avoiding conditionally declared variables
altogether
* PCI passthrough: fix setting VGA to 'none' when marking passed-through
device as 'Primary GPU'
-- Proxmox Support Team <support@proxmox.com> Mon, 19 Oct 2020 15:51:48 +0200
qemu-server (6.2-15) pve; urgency=medium
* fix #2570: add 'keephugepages' config property
* vzdump: log 'finishing' state for Proxmox Backup Server jobs, to avoid
suggesting that the backup is stuck at 100%. This can happen when the
validation and mark of pre-existing chunks needs a bit longer.
-- Proxmox Support Team <support@proxmox.com> Tue, 29 Sep 2020 17:44:28 +0200
qemu-server (6.2-14) pve; urgency=medium
* vzdump: allow bandwidth limit also PBS backup
* avoid a warning when checking the VMs bios
* fix #2862: properly backup (all) VM templates
-- Proxmox Support Team <support@proxmox.com> Mon, 24 Aug 2020 19:33:54 +0200
qemu-server (6.2-13) pve; urgency=medium
* fix use of bandwidth limits with offline storage migrate
* allow one to add CPU features with a dot, like "+sse4.2", correctly
* vzdump: improve logging output and report dirty bitmap state for each disk
* vzdump: display actually uploaded chunks as 'write' speed to conform more
closely with the actual network transmission line-speed.
* fix #2749: vga: disable the display EDID information for the combination of
Windows SeaBIOS and VGA guests to avoid a reduced list of possible screen
resolutions. (Windows may cache the list of possible resolutions,
uninstalling 'Microsoft Basic Display Adapter' and rebooting may then help)
-- Proxmox Support Team <support@proxmox.com> Thu, 20 Aug 2020 11:42:47 +0200
qemu-server (6.2-11) pve; urgency=medium
* fix #2857: restore: pass keyfile to pbs-restore
* fix #2728: die/warn if target is not a replication target when
live-migrating
-- Proxmox Support Team <support@proxmox.com> Fri, 24 Jul 2020 08:13:29 +0200
qemu-server (6.2-10) pve; urgency=medium
* pass-through: fix mdev cmdline generation
* docs: add man page cpu-models.conf(5)
* start: set resume parameter for VM start anytime there is a 'vmstate' in
the config, not just when it has the 'suspend' lock
-- Proxmox Support Team <support@proxmox.com> Mon, 13 Jul 2020 13:37:37 +0200
qemu-server (6.2-9) pve; urgency=medium
* support encrypted pbs backups
-- Proxmox Support Team <support@proxmox.com> Fri, 10 Jul 2020 14:23:46 +0200
qemu-server (6.2-8) pve; urgency=medium
* backup: detect PBS features and use only supported
* fix #2671: include CPU format in man page again
-- Proxmox Support Team <support@proxmox.com> Thu, 09 Jul 2020 15:14:36 +0200
qemu-server (6.2-6) pve; urgency=medium
* vzdump: fix variable redeclaration warning
* make backup log more friendlier to read for humans
-- Proxmox Support Team <support@proxmox.com> Tue, 07 Jul 2020 19:00:09 +0200
qemu-server (6.2-4) pve; urgency=medium
* fix #2787: properly parse vga for vncproxy
* vncproxy: allow to request a generated VNC password
* fix #2794: allow legacy IGD passthrough
* avoid backup command timeout with PBS
* fix #2741: add VM.Config.Cloudinit permission
* enable dirty-bitmap incremental backups for PBS
-- Proxmox Support Team <support@proxmox.com> Tue, 30 Jun 2020 11:33:35 +0200
qemu-server (6.2-3) pve; urgency=medium
* fix #2748: make order of cloudinit interfaces consistent
* fix #2774: add early check for non-managed volumes
* allow to force MTU for a VM net-device
-- Proxmox Support Team <support@proxmox.com> Thu, 04 Jun 2020 11:17:09 +0200
qemu-server (6.2-2) pve; urgency=medium
* adapt net-device hotplug to more strict QMP schema of QEMU 5.0
-- Proxmox Support Team <support@proxmox.com> Fri, 08 May 2020 13:00:18 +0200
qemu-server (6.2-1) pve; urgency=medium
* qmrestore: fix VMA restore from STDIN
-- Proxmox Support Team <support@proxmox.com> Thu, 07 May 2020 21:51:01 +0200
qemu-server (6.1-20) pve; urgency=medium
* cfg2cmd: fix uninitialized value warning on OVMF w/o efidisk
* vzdump: fix backup of templates with stdout as output
* cfg2cmd: set audiodev parameter only on qemu >= 4.2
* api: allow listing custom and default CPU models
-- Proxmox Support Team <support@proxmox.com> Wed, 06 May 2020 17:16:56 +0200
qemu-server (6.1-19) pve; urgency=medium
* clone: use new config_lock_shared
-- Proxmox Support Team <support@proxmox.com> Tue, 5 May 2020 11:22:04 +0200
qemu-server (6.1-18) pve; urgency=medium
* vzdump: assemble: fix skipping all pending and snapshot config entries
* api/destroy: repeat early checks after locking
* migrate: skip rescan for efidisk and shared volumes
-- Proxmox Support Team <support@proxmox.com> Mon, 04 May 2020 17:36:40 +0200
qemu-server (6.1-17) pve; urgency=medium
* backup: never try to freeze in stop mode backup
* Fix #2124: Add support for zstd
-- Proxmox Support Team <support@proxmox.com> Mon, 4 May 2020 14:11:50 +0200
qemu-server (6.1-16) pve; urgency=medium
* spice audio: improve compatibility with QEMU versions newer than 4.2
* migrate: workaround issues with format switch on storage live migration
* fix live migration with replicated unused volumes
* importovf: improve compatibility with OVF files without default namespaces
* backup restore: use correct storage for format check for cloudinit drives
* handle stopping the storage migration NBD server better
-- Proxmox Support Team <support@proxmox.com> Wed, 29 Apr 2020 16:23:24 +0200
qemu-server (6.1-14) pve; urgency=medium
* Use foreach_volume instead of foreach_drive
* Use new storage_migrate interface
* migrate: update config with changed volume IDs
* migrate: allow specifying targetstorage for offline migration
* migrate: sync_disks: use allow_rename to avoid collisions on the target
storage
* migrate: sync_disks: log output of storage_migrate
* migrate: also cleanup disks migrated by storage_migrate in case of failure
-- Proxmox Support Team <support@proxmox.com> Thu, 9 Apr 2020 08:56:44 +0200
qemu-server (6.1-13) pve; urgency=medium
* rescan: fix call to foreach_volume
* migration: fix downtime limit auto-increase
* migrate: allow arbitrary source->target storage maps
* migrate: always check storage permissions and content type
* Include full KVM/QEMU "-cpu" parameter with live-migration and
snapshots/suspend to allow supporting custom CPU models
* fix #2318: allow phys-bits CPU setting
* allow custom CPU models
* config: harmonize bridge pattern to match the same limits of containers
* cpu config: add upcoming EPYC-Rome CPU type
-- Proxmox Support Team <support@proxmox.com> Wed, 08 Apr 2020 17:08:13 +0200
qemu-server (6.1-12) pve; urgency=medium
* CPUConfig: fix module load when pmxcfs is unavailable
* migrate: fix replication false-positives
* migrate: cleanup disk/bitmaps if 'qm start' failed
* migration with targetstorage: check if target storage supports images
* fix efidisks on storages with minimum sizes bigger than OVMF_VARS.fd
* Implement volume-related helpers and use new foreach_volume
-- Proxmox Support Team <support@proxmox.com> Mon, 30 Mar 2020 10:00:13 +0200
qemu-server (6.1-11) pve; urgency=medium
* vzdump: fix regression in backups for specific storage
* custom CPU models: add initial parser and verifier
-- Proxmox Support Team <support@proxmox.com> Thu, 26 Mar 2020 09:00:24 +0100
qemu-server (6.1-10) pve; urgency=medium
* version_guard: early out when major/minor version is high enough
* drive-mirror: add support for incremental sync
* migrate: add replication info to disk overview
* migrate: add live-migration of replicated disks
-- Proxmox Support Team <support@proxmox.com> Wed, 25 Mar 2020 15:16:21 +0100
qemu-server (6.1-8) pve; urgency=medium
* cloudinit: make genisoimage quieter, only output errors
* Append newline to all QGA commands for compatibility with non standard
conforming Apple based guest agent implementation
* add experimental support for proxmox backup server
* fix #2580: api/delete: drop VM from HA resources if purge is set
* improve drive mirror completion over NBD during migration
* add secured unix socket support for NBD storage migration
* Disable memory hotplug for custom NUMA topologies and die on misaligned
memory for hotplug
-- Proxmox Support Team <support@proxmox.com> Fri, 20 Mar 2020 11:11:31 +0100
qemu-server (6.1-7) pve; urgency=medium
* vzdump: always exclude efidisks from backups of machines currently not set
to use OVMF (UEFI)
* Simplify QEMU version check and require at least 3.0+
* Align size to 1 KiB bytes before doing 'qmp block_resize'
* fix #2611: use correct operation when calculating the migration bandwidth
limit
* fix #2612: allow input-data in guest exec and make command optional
* cpu models: add icelake-server and icelake-client
* already add models from future QEMU 4.2 release
* fix #2264: allow one to add a virtio-rng device for improved entropy
bandwidth in a VM
* update_disksize: also update disk size if there was no old size at all
-- Proxmox Support Team <support@proxmox.com> Mon, 09 Mar 2020 19:12:16 +0100
qemu-server (6.1-6) pve; urgency=medium
* allow reading snapshot config for users with VM.Audit on a guest
* fix #2566: increase scsi limit to 31
* fix #2578: check if $target is provided in clone
* update QMP commands to reflect (future) depreacations and changes in QEMU
* resize volume: always request new size from storage after resizing
-- Proxmox Support Team <support@proxmox.com> Mon, 10 Feb 2020 06:40:43 +0100
qemu-server (6.1-5) pve; urgency=medium
* Add QEMU CPU flag querying helpers
* hotplug_pending: remove redundant write/load config calls
* api: vm clone: unlink zombie target VM and firewall config at end of error
cleanup
* add timeout parameter to vm start API/CLI endpoint
* fix #2070: vm_start: for a migrating VM, use current format of disk if
possible
* hotplug_pending: make 'ssd' option non-hotpluggable, it cannot be changed
live on a plugged disk.
-- Proxmox Support Team <support@proxmox.com> Thu, 30 Jan 2020 10:27:33 +0100
qemu-server (6.1-4) pve; urgency=medium
* check if QEMU version is recent enough for requested machine type
* suspend to disk: check and enforce more strict permissions
* update disk size before local disk migration
* hide very long commandline on vm_start/migrate failure
* fix #2493: show QEMU errors in migration log
* api/restore: do not trigger autostart-after-restored task from locked
context
-- Proxmox Support Team <support@proxmox.com> Mon, 16 Dec 2019 16:03:25 +0100
qemu-server (6.1-3) pve; urgency=medium
* create efidisk: poll the real size after volume creation, as some storages
need to create bigger volumes as requested, to cope with their internal
alignment requirements.
* fix #2469: fix qemu-img convert src_format detection, wrongly reverted.
* fix #2510: hostpci: always check if specified device exists
-- Proxmox Support Team <support@proxmox.com> Mon, 09 Dec 2019 11:44:14 +0100
qemu-server (6.1-2) pve; urgency=medium
* api: allow one to remove (hibernation) vmstate
* vzdump: log QGA FS freeze/thaw tries in task log
* skip efidisk0 in hotplug
-- Proxmox Support Team <support@proxmox.com> Sat, 30 Nov 2019 18:38:36 +0100
qemu-server (6.1-1) pve; urgency=medium
* fix #2367: do not allow snapshot with name PENDING
* fix #2469: fix qemu-img convert src_format detection
* implement PVE Version addition for QEMU machine allowing to introduce new
features while keeping migration compatibility more easily
-- Proxmox Support Team <support@proxmox.com> Tue, 26 Nov 2019 13:06:21 +0100
qemu-server (6.0-17) pve; urgency=medium
* PCI(e) pass-through: ensure we fallback to the previous default "0000"
domain again.
-- Proxmox Support Team <support@proxmox.com> Sat, 23 Nov 2019 09:52:09 +0100
qemu-server (6.0-16) pve; urgency=medium
* fix #2473: use of unitialized value
-- Proxmox Support Team <support@proxmox.com> Fri, 22 Nov 2019 14:18:58 +0100
qemu-server (6.0-15) pve; urgency=medium
* api/migration: fix autocomplete for targetstorage
* add 'type' to guest agent format, allowing one to choose between VirtIO
(default) and ISA
* clone: pre-allocate cloud-init disk for destination
* SPICE/QXL: tell Linux VMs that they can add up to 4 display when running
with qemu 4.1 or newer
* add support to tell showcmd helper to assume a specific forced machine
version when assembling a command
* refactor QemuServer to avoid cyclic module dependencies
* fix #2436: pci: do not hardcode pci domain to 0000
* add 'tags' config option for adding meta information to a VM
-- Proxmox Support Team <support@proxmox.com> Wed, 20 Nov 2019 19:41:01 +0100
qemu-server (6.0-14) pve; urgency=medium
* use PVE::DataCenterConfig, use PVE::SSHInfo, use PVE::RRD for RRD data
* refactor migration IP retrieval
* add missing packages to (build-)dependencies
* fix #2457: ga: set-user-password: increase maxLength of password
* fix restoring old VM backups made with Promxox VE earlier than 2.3
* improve test mocking
-- Proxmox Support Team <support@proxmox.com> Mon, 18 Nov 2019 12:12:03 +0100
qemu-server (6.0-13) pve; urgency=medium
* fix #2434: extend machine regex to support stable release machine updates
* prepare to fix #2408, #2355, #2380: use scsi-hd backend for iSCSI as well
* fix deleting pending changes for not yet existing options
* improve hugepage memory size detection
* avoid a race for VMID reservation when importing an OVF manifest to a new
VM
* cleanup importidsk CLI command, and say to which exact disk we imported
* add simple runtime heuristic for IOThread backup support, to ensure the VM
to backup was started with a recent QEMU version.
* QMPClient: ensure QMP connection is also closed in certain edge cases
-- Proxmox Support Team <support@proxmox.com> Wed, 30 Oct 2019 17:43:41 +0100
qemu-server (6.0-12) pve; urgency=medium
* fix regression from 6.0-10 with vmstate restore on RBD
-- Proxmox Support Team <support@proxmox.com> Tue, 22 Oct 2019 16:31:46 +0200
qemu-server (6.0-11) pve; urgency=medium
* fix #1071: VMs with IOThread enabled disks can now be backed up
* fix regression from 6.0-10 with snapshot restore and statefile
* fix regression with from 6.0-10 where forced was always assumed to be true
for applying pending changes
-- Proxmox Support Team <support@proxmox.com> Tue, 22 Oct 2019 12:50:18 +0200
qemu-server (6.0-10) pve; urgency=medium
* fix #2344: ignore cloudinit in replication check
* fix #1291: add option purge for vm_destroy api call
* increase code re-use with pve-container for pending changes in
configuration
* fix #2412: only do the final configuration destroy after all the VMs
resources, and references in other configurations like Firewall or resource
pools was successfully cleaned up
* fix #2171: ensure that non filesystem based statefiles get activated on VM
start
* fix #2395: improve QEMU image converter to cope better with pure file based
sources and iSCSI source and destinations
* fix #2402: allow 1GB hugepages if 2MB is unavailable
* qemu 4.0 : add Cascadelake-Server and KnightsMill Intel CPU models
* fix #2217: don't copy cloudinit disk on clone
-- Proxmox Support Team <support@proxmox.com> Fri, 18 Oct 2019 22:04:50 +0200
qemu-server (6.0-9) pve; urgency=medium
* fix issue where a SPICE remote viewer was disconnected during live migration
* Add VM reboot API/CLI integration, allowing to reboot a VM and applying any
pending changes in-between
* CPU flags: allow one to add aes flag
* fix #2263: die on live migration with local cloudinit disk
* fix #2041, #2272: Add Spice enhancements
* Add support for more (up to 16) PCI(e) devices
* usb: Allow one to make SPICE USB ports USB3 capabile
* allow one to use USB3 for SPICE USB ports with VMs started already with
QEMU version 4.0.0, as live-migrations were not possible with this previous
unsupported setup anyway. Live-snapshots from VMs with a SPICE USB device
which was manually set (wasn't possible over Webinterface) to USB3 with a
machine version of 4.0.0, need to remove the "usb3" flag again from the
snapshot config when restoring it.
* rework kvm_user_version cache mechanism
* api: deletion: check also pending values for serial/usb
* migration api: explicitly clear "online" flag if VM is stopped to avoid
issues with storage migrations which are handled different for stopped VMs
* abort resize disk if current size could not be determined
* fix #2382: delete cloudinit disk before restoring
-- Proxmox Support Team <support@proxmox.com> Thu, 26 Sep 2019 12:01:58 +0200
qemu-server (6.0-7) pve; urgency=medium
* ensure new SPICE audio device works also with 'q35' based VMs
-- Proxmox Support Team <support@proxmox.com> Wed, 24 Jul 2019 15:13:35 +0200
qemu-server (6.0-6) pve; urgency=medium
* Add SPICE audio device support
* fix #2275: die on invalid sendkey
* Make sometimes problematic 'hv-tlbflush' and 'hv-evmcs' CPU flags optional
-- Proxmox Support Team <support@proxmox.com> Tue, 23 Jul 2019 18:20:10 +0200
qemu-server (6.0-5) pve; urgency=medium
* do not pass Proxmox VE internal startdate 'now' to QEMU, it does not
understands it
* use new pcie port hardware for 4.0 and newer q35 machine types
-- Proxmox Support Team <support@proxmox.com> Thu, 11 Jul 2019 19:44:28 +0200
# Older entries have been removed from this changelog.
# To read the complete changelog use `apt changelog qemu-server`.