Skip to main content
Version: 3.1

Changelog

3.1.0 - 2026-02-26​


For the full changelogs visit the release page for each CubeCOS component.

CubeCOS​

Features​

  • Rolling upgrade definition: no VMs go into error states #155 (b3e5e67)
  • feat(k3s) remove unused k3s binary due to security issue (7d13782)
 alert​
  • allow arbitrary file names to be set as alert response exec shell names (471101a)
  • move alert event input for exec from stdin to env (82c5ab9)
  • use batch job to run alert event exec (f0d61fe)
  • limit the resource usage of alert resp job run (4bb0975)
  • add ceph usage alert (835c708)
  • remove ceph usage error event (85e414c)
 app​
  • add framework_check_prerequisites, framework_create, and framework_delete in the app cli (61056d6)
  • add framework_check_port_access in the app cli (24078ee)
  • add framework_list in the app cli (bb07762)
 appctl​
  • register appctl to the core modules (ba05e62)
 build​
  • security artifacts of sbom, scan and cosign #494 (c604258)
 ceph​
  • allow multipath fc and iscsi devices exist during ceph bootstrapping (6f2f364)
  • move multipathd bootstrapping from nova to ceph (f77d272)
  • activate lvm osds during bootstrapping (e05a2ee)
  • backup lvm vg infos after adding mpath devices in case of upgrade (f8212ed)
  • add cli > storage > set_force_use_mpath_devices to switch on or off using multipath devices for ceph if external storages exist (a098de2)
  • force to use the form /dev/mapper/[wwid] as mapped device name in case naming changes between reboots (542163e)
 cinder​
  • set hex_config and hex_translate for adding external storages as cinder backends (69cbba3)
  • add hex_sdk cinder_put_model (1ea416d)
  • apply multipathd config changes (2ce43f3)
  • add hex_sdk cinder_get_model (cc74114)
  • add hex_sdk cinder_get_models (f4bfe86)
  • add hex_sdk cinder_put_models (61d7b4b)
  • add hex_sdk cinder_delete_model (4d73a58)
  • add hex_sdk cinder_put_storage (99b90ff)
  • apply configs and set volume type properties in cinder_put_storage (2be2365)
  • catch volume type properties apply error (e551700)
  • add hex_sdk cinder_test_storage (e9275aa)
  • add hex_sdk cinder_get_storages (29d07e7)
  • add hex_sdk cinder_get_storage (1fd0a8b)
  • add hex_sdk cinder_delete_storage (7963cae)
  • add hex_sdk cinder_get_default_storage (ed57507)
  • add hex_sdk cinder_set_default_storage (568c453)
  • add hex_sdk cinder_get_active_multipath_setting (0a279e5)
  • migrate config files for external storages (382e86c)
  • add dell emc powerstore as a builtin model (88c2c2c)
  • add generic nfs as a builtin model (078506b)
  • provide existing volume manage capabilities on NFS volume backends (3c42d66)
  • add cli > iaas > volume > migrate_large_volume_from_nfs (3a6fff6)
  • filter out already managed volumes on nfs backends (4d69c7f)
  • add volume name input to cli > iaas > volume > migrate_large_volume_from_nfs (c7dcc66)
  • add dest volume type input to cli migrate_large_volume_from_nfs (9b89b97)
  • add virt-v2v volume conversion input to cli migrate_large_volume_from_nfs (7490447)
  • add cli > iaas > volume > move to move a volume to a backend (f5e90a1)
  • take over the control of the existing volume (37c1c43)
  • check if the file is in the raw format (fd81337)
  • check if the project quota is enough to manage the volume (55fe568)
  • perform volume conversions using virt-v2v (31a0074)
  • update the target file path for cinder manage to the converted volume file path (49bc3f0)
  • set volume image properties to the volume (a9bf9c9)
  • parse volume metadata (c4873fe)
  • add cli > iaas > volume_backend > list_models (9026daf)
  • add cli > iaas > volume_backend > list_drivers (49c2dea)
  • add cli > iaas > volume_backend > list_model (964d7c2)
  • add cli > iaas > volume_backend > set_model (840ff21)
  • add cli > iaas > volume_backend > delete_model (e52ac24)
  • add cli > iaas > volume_backend > get_active_multipath_setting (50c7fdf)
  • add cli > iaas > volume_backend > get_storages (2e82c35)
  • add cli > iaas > volume_backend > get_storage (5f0821b)
  • add cli > iaas > volume_backend > test_storage (c9a3974)
  • add cli > iaas > volume_backend > get_default_storage (edaadf0)
  • add cli > iaas > volume_backend > set_default_storage (9d22f81)
  • add cli > iaas > volume_backend > delete_storage (7fc8f1a)
  • add cli > iaas > volume_backend > set_storage (31e221a)
  • sync models on control nodes to pure compute nodes (041b554)
  • make virt-v2v to use tmp directories on nfs instead for cli > iaas > volume_backend > manage_existing_from_nfs (73c3926)
  • print preconditions for cli > iaas > volume > manage_existing_from_nfs (4bd0577)
  • add fujitsu eternus dx series fc and iscsi as builtin models (24490de)
 elk​
  • build opensearch-dashboard by ourself to solve security issue (5f50687)
  • Revert build opensearch-dashboard by ourself to solve security issue (46fcd38)
  • security patch for opensearch-dashboard (4342dc7)
 glance​
  • config glance for setting cinder volume-backed store (eb4c7e9)
 gpu​
  • upgrade nvidia vgpu manage driver to v580.105.06 to support rtx pro 6000 blackwell server edition (4cd337f)
  • install fabric manager (0a24d78)
  • back port vfio variant driver framework support for nvidia gpu sr-iov vgpu from openstack epoxy (bcd9449)
  • add cli > gpu > nvlink_enable / nvlink_disable to manage nvlink devices (466bcd1)
 hex_cli​
  • add openstack exec (3705029)
  • add command iaas volume_backend list to list existing and configured backends (7502109)
  • use crontab to schedule jobs for ceph osd removals (93c59ec)
  • avoid scheduling duplicated ceph_osd_safe_remove jobs (8d8b57c)
  • use crontab to schedule jobs for ceph osd disk removals (fe080c6)
  • support mpath devices in cli > storage > list_avail (b14a4f9)
  • disable mpath device support for Ceph if external storage is enable, unless being forced (75068b8)
  • exclude mpath devices in cli > storage > add_disk & add_avail (3b950f9)
  • make ceph_osd_host_remove be compatible with the new ceph_osd_remove logics and prevent cron job from rerunning jobs after updating the cron file (e6c5278)
  • remove checks for duplicated job schedules since underlying functions already perform the check (acaa7bb)
  • make cli > storage > add_disk be able to add mpath devices (e6da13d)
 hex_config​
  • load default configs of cinder and glance for the support team (6ea9308)
  • hide mongodb start up messages (f0fa3ec)
  • migrate needed config files for upgrades (06310fa)
 hex_sdk​
  • get event list for attribute-based alert trigger (41ac2b4)
  • accept json input for alert setting put setting receiver exec shell (6277d6f)
  • allow customized triggers (b1a4304)
  • add alert_delete_trigger to delete a trigger (4671232)
  • provide exec call isolation (17b062d)
  • add json helper functions (96a5bf7)
  • add file system helper functions (923eb0b)
  • stats_partition #448 (73c6b78)
  • add logging functions (d3cc37f)
 jail​
  • mute emulate warning and fix nested build permission issue (534dcdb)
  • enable podman daemon for dapper build process (c244743)
 k3s​
  • build k3s by ourself to solve security issue (415e05c)
 kapacitor​
  • migrate registered trigger exec shells and bins (c9a1495)
  • pour alerts in topic instance-events to topic events (befe34f)
 kernel​
  • move to kernel-ml (temporarily skip compiling nvidia) (13ac3f7)
  • switch to kernel-6.12 to satisfy nvidia driver compilation (a2d795e)
 keycloak​
  • set keycloak admin password (5e93f7e)
  • save keycloak admin password as a k8s secret (79f45b7)
  • add command get_keycloak_admin_password (dc2e58f)
  • update the admin password used in the terraform provider (4a3f5ae)
  • add a setting health check for keycloak admin password not synced (c696001)
  • hide db connection infos as k3s secrets (c67bbb4)
 log​
  • send hex logs to separated files (5e36f43)
 make​
  • generate sha256 checksum (924968d)
  • push checksum computations of terminal artifacts to background and nohup them (c65a354)
  • include the gpu driver in builds (ea271d9)
 manila​
  • config manila for setting cinder volume types of external storages (23f75ae)
 mongodb​
 neutron​
  • enable qos plugin (84bf7c0)
  • enable the subnet_dns_publish_fixed_ip for dns feature (d30b180)
  • publish dns record automatically when setup dns-domain (1828667)
 nova​
  • set vTPM config to nova (cf1a8fa)
 oci image​
  • add a troubleshooting tool by api add busybox-extras (efe5515)
 prometheus​
  • build promtool by ourself to solve security issue (6784fdb)
  • build prometheus and promtool by ourself to solve security issue (20e9024)
 rancher​
  • upgrade rancher server and cli to v2.11.2 to support newer k8s versions (379a5b2)
  • deprecate cube node driver and remove related initial process (f70644b)
  • build rancher by ourself to solve security issue (dc6af23)
 sbom​
  • omit more directories during sbom generations (8ce549c)
  • omit /etc in sbom generations (d76547b)
 security​
  • drop icmp requests and replies (28cf3d8)
 senlin​
  • add cli > management > cleanup > cleanup_senlin to remove openstack senlin remnants (4c5f2b0)
 skyline​
  • update to latest master to include the security fix (fed7c2e)
 telegraf​
  • update mod to solve the security issue (6c7962d)
 terraform​
  • build terraform by ourself to solve security issue (03f52fe)
  • build keycloak provider by ourself to solve security issue (1d9dbb4)
  • update go-getter and x-crypto to solve serurity issue (f8e14f0)
 yq​
  • build yq by ourself to solve security issue (9cedc46)

Bug Fixes​

  • filter out intermediated grep results that pollutes stdout #155 (e1cd557)
  • SDK mgmt_ip doesn't handle bonding + vlan #205 (e3a16e6)
  • SDK GetParentIfname bad inputs, causing malfunction of password_init #155 (4968f27)
  • SDK ceph_osd_list outputs messed up by concurrent instances #225 (b306f55)
  • control-converged pacemaker_remote_add leads to vip loss #155 (a1415fb)
  • bootstrap can fail because of temporary master node VIP #155 (27dc983)
  • compute node fails to init octavia-hm0 #155 (ddcc8b1)
  • rolling-upgrade master control unresponsive ceph mon #155 (b306ee2)
  • rolling-upgrade ensure only master control has vip #155 (a73b62c)
  • proj_functions to use standardized timeout variables (a22901d)
  • rolling-upgrade VM evacuation continues irrespective of failures (5b16df1)
  • bootstrap logic mistakes #155 (0b7c6fa)
  • non-master control should not lose vip (3267bfe)
  • rolling upgrade cluster level SDK pacemaker status #155 (c0a0da3)
  • FTS roll is undef so health_vip_check doesn't work #155 (770c08f)
  • stopping pacemaker with systemd can take 20 min #155 (1fcff37)
  • infinite loop of hex_sdk #155 (5ff2c7f)
  • SDK ceh_mds check output and detail not consistent #380 (2cf6cc3)
  • SDK is_checking bug #155 (599ca6f)
  • fixpack hide unwanted command outputs #86 (5823fbf)
  • rolling-upgrade do not run check_repair which brings up pcs on non-control nodes #155 (e89967e, 401)
  • rolling-upgrade unsuccessful live-migration #155 (41b471c)
  • image import misses progress bar in some cases #426 (d66958f)
 alert​
  • properly escape event inputs for alert response exec job (a534138)
 app frameworks​
  • change the 'rancher-cluster-image.raw' to 'rancher-cluster-image-rke2-v1.32.4.raw' (3dc7d7b)
  • fix the image typo in the extpack's rancher md5 check (f377670)
 appctl​
  • fix the incorrect file preparation path (297d3f6)
  • fix the git path conflict during building artifacts (365f4d5)
  • fix the incorrect var usage in the appctl/copy-images.sh (3547870)
  • fix the incorrect var usage in the appctl/copy-images.sh (fd555e4)
  • fix the duplicated ./source dir during build time (811b2d5)
  • do APPCTL_GIT_DIR removal before creation (d1dae49)
  • use cp -rf to copy the plugin dir (774cfa0)
  • fix the build process in the appctl's Makefile (59a3014)
  • fix the offline image preparation issue (e731ef6)
 appfw​
  • update docker-registry helm chart repo url (9af13a2)
 build​
  • ignore rpm --import Warning: Unsupported version of key: V6 (a44795b)
  • unlock crun-1.20-2.el9 which is unavailable from upstreams (2bdd69c)
  • increase timeout while reducing loop retries (eb2a2e9)
  • bump up to kernel-6.12.69 (0e4a377)
  • java.lang.OutOfMemoryError: Java heap space hitting 1G cap (a74d4a2)
 ceph​
  • check if external storage is set on vip control node (90aba45)
 cinder​
  • set up the service on fts, set the default volume type to the built-in if empty, and check if files exist before touching them (9919d9b)
  • skip empty volume backend name (045959b)
  • stop disabling qemu-guest-agent and iscsi-onboot services (1d176e5)
  • fix type (b4dfaf7)
  • add the missing successful output of hex_sdk cinder_set_default_storage (23a95ca)
  • set the correct file path (060aa1b)
  • restore the built-in model's multipath settings if it exists (0652834)
  • escape the colon in volume type for cinder_get_volume_type_properties (d583712)
  • failed to filter out duplicated storage names (86ad512)
  • fix typos in sdk interfaces (d1042cd)
  • support multiple backends using the same model (fd82fc8)
  • parse ini outputs without any sections (fbe66f7)
  • allow one line command input for cli migrate_large_volume_from_nfs (79c0f83)
  • escape user input values in bash executions (166ce85)
  • always check the project quotas for managing a volume (542a4c5)
  • use absolute paths to call the binaries instead (67d1944)
  • drain the pipe in ExecSync for the missing contents (cb55210)
  • check if storage backend config files are modified or not (e91f7c1)
  • get volume backend pools when backends have similar names (3b7ab87)
  • check volumes from all projects before deleting the external storage (acbcfc9)
 cubectl​
  • fix cubectl's lib conflict issue (01f0399)
  • upgrade the golang lib to 1.24.2 for resoving the net/http golib issue (6a0f87f)
 etcd​
  • upgrade the etcd to 3.5.23 for resoving the net/http golib issue (6d4b8fd)
 fixpack_history​
  • clear history after FW upgrade #695 (7f12024)
 glance​
  • set permissions before opening the file writing stream (8a53609)
  • skip empty store backend name (f697b52)
  • reactivate logging for glance (5bc97b2)
  • allow openstack user cinder with role service to set image locations (7b6bd1e)
  • lift project service quota for volume-backed images (9f21a3a)
 haproxy​
  • separate cookie name token from different backends (96162be)
  • remove the 'replace-header' of cookie for cube-cos-api and ceph-dashboard (2640c05)
 hex_cli​
  • cluster remove_mode does not remove its osds #101 (7317adc)
  • remove the Quiet() for the ceph restful create-key and ceph restful list-keys (120a666)
  • check array index boundaries (f0c155b)
  • use flock to avoid race conditions on read then write crontab (6d34c52)
  • use safer checks to test if a disk is ready to be removed (b5026d3)
  • remove not needed /sys/class/scsi_host/*/scan outputs (c07f850)
 hex_config​
  • use the correct boot status cli command (74a49d4)
  • resolve logrotate config conflict with rsyslog default configs (db3459c)
  • Cirros image fails to import when glance is not fully ready (c55e993)
  • limit file permissions of written files (61d0ed8)
 hex_fixpack_install​
 hex_install​
  • refactor FormatDataHDD and exclude non-writable and zero size devices (2900ffe)
 hex_sdk​
  • check file existence and adjust the wrong input parameter in alert_put_setting_receiver_exec_shell for alert_add_update_setting_receiver_exec_shell (56b43ea)
  • fix typo (b01a852)
  • stop duplicating tuning outputs (4b0bb33)
  • block tcp port 8080 and accept local traffics (95abeaa)
  • avoid having multiple Cirros img in Glance (3eafd3a)
  • CLI image import with progress bar would hang #128 (70e2cb6)
  • image import to check volumes quota limit and space #445 (a740cf8)
  • image import quota check not cover correct proj_name #445 (8e9c494)
  • image import visibility shared should be public #463 (90aa2f9)
  • image import to ext-storage backend via Glance #468 (6032a2a)
  • image import to ext-storage backend not converted via virt-v2v #468 (1286d17)
  • image_import to handle quota -1 value #563 (c081897)
  • ceph_osd_host_remove never finishes #679 (84fb304)
  • trigger migrate_fixpack #695 (3593507)
  • create s3 log bucket during cluster_start #703 (8d957da)
 httpd​
  • only allow control management ips to access apache mod_status /server-status page (b0fd95a)
 install​
  • exclude fc and iscsi hosts from the install process (fba61cf)
 jail​
  • upgrade the golang lib to 1.24.2 for resoving the net/http golib issue (d1d36fd)
  • correct the typo in parameter pass (2458402)
 k3s​
  • adjust the process flow for generating the ceph rgw credential (190c93e)
  • upgrade the k3s from 1.26.6 to 1.32.11 for resoving the net/http golib issue (90d60e5)
  • Revert upgrade the k3s from 1.26.6 to 1.32.11 for resoving the net/http golib issue (b41367f)
  • correct the final build k3s binary to fix kubectl not found error (2b2c7c6)
  • handle the dependency issue of klog after upgrade to 1.24 (302ad98)
  • wait for metrics-server pod and api service after k3s boostrapping (67a3681)
  • correct bash syntax error while waiting for metrics-server api service (64e7f1c)
 kapacitor​
  • set the correct event messages for ceph usage alerts (22294a9)
  • adjust the event list per ceph usage alert threshold change (b91f0a0)
  • filter out illegal characters in event handler names (40ca7ec)
 kernel​
  • bump kernel version from 6.1.136 to 6.1.149 (eec2560)
 keycloak​
  • avoid updating keycloak during rolling upgrade (fbd247e)
  • use the correct status flag (59f68b4)
  • deploy at least one pod of keycloak (fdb9fd0)
  • repair 0 replica in the helm release (eb66edd)
  • fix the typo in the keycloak makefile (7ef8b08)
  • ensure k3s is running before deploying keycloak (a88de9d)
  • hide error messages from kubectl and helm cli (0981800)
  • set required file permissions on keycloak idp xml file for keystone mellon to work (bde04a4)
 linux​
  • fix linux user group database misconfigurations (2c04739)
 make​
  • add the missing chmod for sha256 of pkg parts (a164315)
  • preserve PROJ_PPUISO proj.pkgiso builds (8f11d86)
  • remove the sha256 checksum files from the previous build (4c887fd)
  • use find instead since extended globbing is not supported in Bourne shell (fddaf1b)
  • avoid using subshell after nohup (eb045c8)
  • disambiguate semicolon and ampersand (e0d3a8f)
  • bump qemu-kvm version to 17:9.1.0-26.el9 (a7b9c7f)
  • bump qemu-kvm version to 17:9.1.0-27.el9 (1969f4d)
  • downgrade qemu-kvm version to 17:9.1.0-24.el9 (151c173)
  • place yq to the correct location (6cf33ce)
 mongodb​
  • upgrade mongodb from 7.0.25 to 7.0.28 to fix the CVE-2025-14847 (4d0d0d6)
 non-rolling​
  • pcs has no vip resource #155 (26639cf)
 nova​
  • allow volume-backed instance snapshots to be used to create vms (cbf03b7)
 pcs​
  • conflicting constraints (517c2a8)
 rancher​
  • fix the unexpected rancher pod pending during version upgrade in 1cc and 1ec (b9c64a7)
  • upgrade the terraform rancher module to 7.0.0 to support the rancher server(v2.11.2) operation (f86ccd0)
 rc.nicdetect.sh​
  • handle network multiplexer for physical and VM #517 (0f61538)
 rolling​
  • stale mariadbd hinders service to start #155 (4261234)
  • mariadbd failed to be repaired #155 (480231f)
 rolling-upgrade​
  • keep vip on old firmware to minimize VM down time #155 (d83ec8d)
  • temporarily remove vaw for backward compatibilities #155 (8ce11b2)
 s3​
  • empty files in s3://log bucket (28c0850)
 security​
  • monasca log4j and misc pip pkgs (5d2e77a)
  • bump up ELK stack versions (0af3d28)
 terraform​
  • upgrade the terraform from 0.14.3 to 1.12.0 for resoving the net/http golib issue (2d282cf)
  • Revert upgrade the terraform from 0.14.3 to 1.12.0 for resoving the net/http golib issue (cbfdbab)
  • skip sum check for self build rancher provider (fd8e2e7)
 watcher​
  • patch watcher's workload balancing strategy (32d12a6)
  • add CONFIG_OBSERVES keystone in the watcher (5dc24c0)
  • remove the unnecessary vm_workload_consolidation files (1d3f139)
  • fix the incorrect statistic unit for instance_ram_usage from GiB to MiB (6cc93e4)

Components Updated​

  • CentOS Stream 9
  • Kernel: 6.12.69
  • OVN: 23.03.1
  • Ceph: 17.2.6 (Quincy)
  • Rancher: v2.11.2
  • Nvidia GPU driver and Fabric Manager: v580.105.06

CubeCOS UI​

Full changelog

CubeCOS API​

Full changelog

3.0.0 - 2025-05-27​


For the full changelogs visit the release page for each CubeCOS component.

CubeCOS​

Added​

Alert​

  • Add the alert setting policy and its CLI, translate, and hex_sdk functions
  • Add the alert trigger policy and its CLI
  • Add hex_sdk functions to put trigger and delete on cascade from alert setting to alert trigger
  • Add the alert trigger translate and hex_sdk functions and adjust hex_config Kapacitor for alert setting and alert trigger
  • Remove the old cli command configure
  • Migrate alert_resp1_0 policy to alert_setting1_0 and alert_resp2_0 policy

Deployment​

  • Set configs for hosting UI and proxying requests to API

LMI​

  • Remove LMI, UI and API to take over ports and serve rancher node drivers through httpd

Miscellaneous​

  • Json output for hex_config -P and regex validate for str type
  • SDK -f json DumpInterface (#17)
  • adjust configurations of Kapacitor, haproxy, keycloak, and API
  • Tuning_dump to display net interfaces ethx according to setting
  • CLI image import with additional properties
  • SDK opensearch_ops_reqid_url req-id (#54)

Security​

  • Set up authentication for MongoDB and only expose MongoDB and InfluxDB on MGNT IP and VIP

Changed​

Image Services​

  • Pull the fix for image import as a Cinder volume

Login​

  • Pull login greeting messages on Keycloak login page from the old repo

MongoDB​

  • Pull MongoDB service set up from the old repo

API​

  • No longer manage configurations for API

Fixed​

AppFramework​

  • Revert changes regarding the node driver name

hex_cli​

  • Remove the duplicated output line in image > import

hex_sdk​

  • Open output from hex_sdk os_nova_instance_ping to telegraf for grafana vm instance ping graph

License​

  • Use UTC time to read the expiry date from the license

Miscellaneous​

  • hex_config sdk_run should return what SDK returns (#29)
  • Fix API health check & repair logics, add name for license, use role admin in diagnostics, and fix word splitting of image name
  • SDK -f json DumpInterface to support bonding interfaces (#60)
  • VMs in error state after cluster powercycle (#71)
  • SDK tuning_dump net.if.mtu not expanded (#76)
  • Bootstrapping in parallel can fail due to vip change (#79)
  • SDK MongoDB_check syntax errors
  • cmd CEPH variable not expanded
  • SDK with db logging and other capabilities
  • Cron job to purge old data in object store
  • SDK opensearch_ops_reqid_url reqid last (default is last 7 days)
  • Rolling upgrade enhancements
  • Unable to open unix socket on ext4 with kernel > 6.1.123
  • Include the full path for path rewrite in skyline-console nginx proxy to cube-cos-API

Components Updated​

  • CentOS Stream 9
  • Kernel: 6.1.136-1
  • OVN: 23.03.1
  • Ceph: 17.2.6 (Quincy)
  • Rancher: v2.7.9
  • Nvidia GPU driver: 535.104

Compatibility​

  • Compatible with CubeCMP 2.0 version.
  • Supports CubeVDI Driver and integration.

Known Issues​

None.


hex​

Added​

alert​

  • Add the alert setting policy and its CLI, translate, and hex_sdk functions
  • Add the alert trigger policy and its CLI
  • Hex_config for alert setting and alert trigger

Changed​

  • Update the banner theme

Miscellaneous​

  • Normalize shell coding styles
  • Json output for hex_config -P and regex validate for str type
  • SDK -f json DumpInterface
  • Unpublish wildcard tunings
  • Unpublish tuning time.timezone
  • Upgrade GRUB2_VER to 2.06-104

hex_cli​

  • Add command tree to list out all enabled commands

license​

  • Add name and rename SLA to support plan

Fixed​

license​

  • Use UTC time to read the expiry date from the license

Miscellaneous​

  • SDK -f json DumpInterfaces to support bodning info
  • More readable msg when interface properties are not detected
  • SDK tuning_dump net.if.mtu not expanded
  • SDK tuning_dump not include bonding interfaces

CubeCOS UI​

Full changelog

CubeCOS API​

Full changelog


2.5.0​

Added​

  • Introduced disk encryption for secure disk management; each drive now defaults to one OSD.
  • Added login greeting message feature for user-facing alerts on login.
  • Integrated security scanning into the build pipeline to block high-severity vulnerabilities from untrusted packages.
  • Enabled file system integrity monitoring with auto-restore on reboot.
  • Introduced USB storage device blocking by default, with CLI-based control via cubesys.probeusb.

Changed​

  • Enhanced the Remove_disk CLI command with a safe/force option via an index menu.

Fixed​

  • Improved output from the list_osd command to provide more detailed storage status information.

Components Updated​

  • CentOS Stream 9 base OS
  • Kernel: 5.14.0-435
  • Nova: 25.3.0 (Yoga)
  • OVN: 23.03
  • Ceph: 17.2.6 (Quincy)
  • Rancher: v2.7.9
  • Nvidia GPU driver: 535.104

Compatibility​

  • Compatible with CubeCMP 1.7 or later versions.
  • Supports CubeVDI Driver and integration (requires CubeCOS version 2.4 or later with CubeVDI Essentials).

Known Issues​

  • OTP login causes Keycloak login page rendering issue when enabled.