proxmox-datacenter-manager (1.1.4) trixie; urgency=medium * remote PBS connections: use MultiClient which has timeout support, to avoid stalling metric collection. * metric collection: drop datapoints more than a minute ahead of the request time, to avoid large gaps in metrics in case remote clock jumped ahead. * remote connections: only fail over client errors for idempotent methods, but keep fail over on Connect errors for all methods. * support status: always fetch fresh subscription information from remote candidates on update, the cache lacks required information. * cache: return early instead of doing an unnecessary cache lookup if queried with special max age value 0 -- Proxmox Support Team Wed, 03 Jun 2026 14:40:32 +0200 proxmox-datacenter-manager (1.1.3) trixie; urgency=medium * remote connections: fail over to another cluster node again when a node fails to connect (DNS, TCP, or TLS handshake); such failures had stopped triggering failover, so a transient issue on one node could fail a request that another node could have served. * remote connections: retry a transient connect failure on a node once per request - for a single-node remote its only recovery, and for a cluster a best-effort retry of the first node that runs only after every other node was tried, so it never delays failover. -- Proxmox Support Team Sat, 30 May 2026 00:29:22 +0200 proxmox-datacenter-manager (1.1.2) trixie; urgency=medium * auto-installer: reject a rendered host address (CIDR) that carries no netmask, which the parser would otherwise silently treat as a single-host /32 (or /128), yielding a misconfigured network for a templated address. * cli: support listing the guests of multiple remotes in a single invocation. * docs: recommend the dedicated container meta-package for container installations, so installing does not pull in a host kernel. -- Proxmox Support Team Fri, 29 May 2026 22:22:09 +0200 proxmox-datacenter-manager (1.1.1) trixie; urgency=medium * api: certificates: let any authenticated user query the host's TLS certificate information, which every TLS connection exposes anyway, and add a matching pdm-client method, so the web interface can prefill the fingerprint for the automated-installer preparation command. * auto-installer: - match prepared answers in the configuration's stable order and let the first match win, rather than a random match that could change across restarts when several answers overlap; warn when a further answer also matches, so overlapping target filters can be disambiguated - reject negative template counter values at the API, not only malformed counter names, as a counter only ever increments from its seed - name the offending field and include the rendered value when a templated CIDR, gateway, or DNS address renders to an invalid value, instead of a bare "invalid IP address" * map: clear a remote's cached location once its last location is removed, so clients are no longer served a stale location depending on the requested cache max age. * docs: add guides for Ceph cluster monitoring and central guest management, document the optional per-answer auto-installer subscription key, and clarify where remote map locations are configured. -- Proxmox Support Team Thu, 28 May 2026 00:01:11 +0200 proxmox-datacenter-manager (1.1.0) trixie; urgency=medium * bump for 1.1 point release. -- Proxmox Support Team Wed, 27 May 2026 02:26:41 +0200 proxmox-datacenter-manager (1.0.7) trixie; urgency=medium * cli: improve fido2 authentication when multiple security keys are connected: ask the user to touch the device they want to use instead of trying them one after the other, and print the device info to stdout so it is visible which device's PIN to enter. * api: auto-installer: propagate the proxmox-wide token-name schema to the answer-token list, create and update endpoints, so malformed token IDs are rejected at the API boundary instead of being accepted and failing later. * auto-installer: store the prepared-answer cidr, gateway and dns fields as strings so they can hold a MiniJinja template expression (e.g. an address with a {{ installation_nr }} counter); the rendered answer is still parsed into typed addresses, so an invalid result is rejected at render time. -- Proxmox Support Team Tue, 26 May 2026 19:00:00 +0200 proxmox-datacenter-manager (1.0.6) trixie; urgency=medium * add central guest snapshot management for QEMU and LXC guests: list, create, roll back, delete and edit the description of a snapshot through new API endpoints and CLI subcommands. * remotes: add an endpoint and CLI subcommands to re-probe a remote node's TLS certificate and update or clear the stored fingerprint, so a rotated certificate no longer forces removing and re-adding the remote. This is a stop-gap until all Proxmox products support staged certificate rotation. * auto-installer: support an optional Proxmox subscription key per prepared answer, forwarded into the rendered answer file so an automated install can register the new node's subscription without an operator step. * subscription: tolerate a corrupt shadow file when removing a pool key. Drop the cached shadow blob best-effort after the authoritative pool config removal succeeds, instead of refusing the removal up front. * docs: document recovery from a rotated TLS certificate on a remote: when the stored fingerprint is pinned, how the mismatch surfaces as an opaque connection error, and how to re-probe and update or clear the pin. * docs: refresh the subscription registry screenshots and notes. -- Proxmox Support Team Tue, 26 May 2026 03:08:23 +0200 proxmox-datacenter-manager (1.0.5) trixie; urgency=medium * add read-only Ceph cluster monitoring for Proxmox VE remotes: auto-detect Ceph clusters, track their state with a periodic detection sweep, and expose cluster status, OSDs, monitors, managers, metadata servers, CephFS instances and pools through new API endpoints, cached via the shared API cache. * add backend support for a geographic world map of remotes: serve a geojson world map and a cached per-node location-info API that tolerates unreachable remotes and nodes. * add subscription key pool management: store subscription keys centrally and assign them to remotes, with API endpoints, admin CLI subcommands, and actions to check, adopt and clear keys, including per-node revert. * api: add an endpoint to resume paused or suspended QEMU guests on a remote, complementing the existing start, stop and shutdown actions. * views api: return a configuration digest when retrieving a view, so clients can detect and avoid clobbering concurrent edits; also run the view handlers in the proxy process and improve schema and error messages. * rework resource, subscription and remote-update caching onto a new shared, persistent key-value cache, making cached data self-healing and more consistent across these subsystems. * api: require only the Audit privilege, instead of Modify, to list certificate information. * docs: document the subscription key pool registry and the available custom-view widgets, including the new map widget. * backend package: recommend ifupdown2. -- Proxmox Support Team Sun, 24 May 2026 04:14:27 +0200 proxmox-datacenter-manager (1.0.4) trixie; urgency=medium * admin cli: add commands to manage ACME certificates, accounts and plugins. * acme: improve interoperability with various ACME server implementations: accept HTTP 204 responses to nonce requests (#6939), treat non-required URL fields in directory responses as optional, and disable accidental base64 padding in request payloads. * views api: return a clearer error message if a view layout contains any unknown widgets. * add metric collection system for local metrics. Every 10 seconds, this new subsystem collects various system metrics and stores them in a local database. * api: drop unused, broken /nodes/localhost/rrdata API handler. It always failed due to a missing 'node' parameter in the definition. Also, there was a typo in the path ('rrdata' instead of 'rrddata'). * api: add /nodes/localhost/rrddata route, returning metrics for the local PDM node, to be consumed by the RRD graphs in the UI. * api: subscription status cache: ensure that a request with max_age=0 actually triggers a fresh fetch instead of returning a same-second cache hit, restoring the freshness guarantee for direct ?max-age=0 callers and the uncached status collector. * fix #5076: openid realms: support an 'audiences' property to declare additional trusted audiences beyond the configured client-id, improving compatibility with providers that issue tokens carrying multiple audiences. * automated installer integration: PDM can now serve answer files via HTTPS for the Proxmox auto-installer. Operators prepare per-product answer configurations, restrict retrieval through bearer tokens, and view past and in-flight installations from a new "Automated Installations" tab in the Remotes menu. Permissions are scoped under a new /system/auto-installation ACL path. * auto-installer: templated fields in the answer file (FQDN, mailto, network addresses) are evaluated through MiniJinja over the requesting machine's system information, and the post-installation hook callback is authenticated with a per-installation secret PDM hands out alongside the answer. -- Proxmox Support Team Fri, 15 May 2026 11:09:28 +0200 proxmox-datacenter-manager (1.0.3) trixie; urgency=medium * client cli: support the 'target-vmid' parameter for the 'pve qemu remote-migrate' and 'pve lxc remote-migrate' commands. * validate a remote node's URL when updating a remote config. * docs: ZFS: repair broken formatting and other minor cosmetic touch-ups. * docs: ZFS: fix instructions on how to change the kernel's swappiness via a drop-in file in /etc/sysctl.d/ on Debian Trixie. * docs: fix typo in FAQ section. * client/admin cli: fix shell completion for bash by providing completion files with the expected file extension. * improve error reporting for the backup snapshots API endpoint for PBS remotes by passing through the original error from the PBS API. * fix #7120: remote updates: drop vanished nodes and remotes from the update cache on full refresh, avoiding stale entries after cluster topology changes. * cli: fix zsh shell completion definitions. * api: pbs: add realms endpoint to enumerate available authentication domains on PBS remotes. * fix #6914: add option to delete the API token on the remote when removing a remote, avoiding errors when re-adding the same remote later. Supported in the API, client CLI, and admin CLI. * remote tasks: add API endpoint to manually trigger a task cache refresh for a single or all remotes. -- Proxmox Support Team Tue, 10 Mar 2026 00:06:23 +0100 proxmox-datacenter-manager (1.0.2) trixie; urgency=medium * admin cli: add `support-status` sub-command * server: include the new `support-status` sub-command in the system report. * update proxmox-apt and related api-types for more robust package name handling. -- Proxmox Support Team Mon, 22 Dec 2025 17:01:33 +0100 proxmox-datacenter-manager (1.0.1) trixie; urgency=medium * api: subscription check avoid potential division through zero and improve message. * docs: roadmap: typo and phrasing fixes. * api: subscription check: fix type comparison and improve log message for better introspection. -- Proxmox Support Team Thu, 04 Dec 2025 15:19:35 +0100 proxmox-datacenter-manager (1.0.0) trixie; urgency=medium * bump for first 1.0 stable version. * server: move ldap passwords file inside the access folder. Automatically migrate existing files on upgrade. * daily update: wire up subscription handling. * docs: include views.cfg options and add man-page. * docs: add basic chapters for Views, Access Control, SDN and extend the one for the web-based user interface. * docs: add some basic screenshots from the web ui. -- Proxmox Support Team Wed, 03 Dec 2025 19:36:13 +0100 proxmox-datacenter-manager (0.9.9) trixie; urgency=medium * pve: add api call to get the cached version info from remotes. * api structure: move /remotes/{id} to /remotes/remote/{id} * api: move remote-tasks, remote-updates and metrics-collection under /remotes, as these do not warrant to be a top-level API endpoint. * admin cli: add command for getting all remote subscriptions status. * system report: include subscriptions status of all remotes. * resource: fix storage identifier. * docs: fix various typos and wording issues all over the place. * docs: add basic roadmap including release history. -- Proxmox Support Team Tue, 02 Dec 2025 19:25:57 +0100 proxmox-datacenter-manager (0.9.8) trixie; urgency=medium * add initial built-in documentation. * move grub snippet over to new proxmox-datacenter-manager-meta package. * api: apt versions: track new proxmox-datacenter-manager-meta as virtual package. * remote.cfg: clean up shadow entries for non-existent remotes. -- Proxmox Support Team Tue, 02 Dec 2025 09:52:53 +0100 proxmox-datacenter-manager (0.9.7) trixie; urgency=medium * remote config: use shadow config to safe token secrets to reduce the likelihood that an user leaks theirs by accident when sharing the remote config without knowing the full implications. * add basic system-report API endpoint and admin CLI command. * add an opt-in api-viewer package. * api: add subscription endpoints for PDM itself. There is no dedicated subscription, but existing customers with an active Basic or higher subscription for their Proxmox remotes also gain access to the Proxmox Datacenter Manager Enterprise-Repository and support. For this to work 90% of the remotes need to have a subscription, and out of those less than 40% can be community ones. * api: sdn: add endpoints for ip-vrf and mac-vrf. * priv daemon: create jobstate directory after the state dir. * api: sdn: recognize pending status for networks to avoid that they are shown as being in an unknown status. * api: add endpoints for PVE node config and cluster options to be able to render the remote notes. -- Proxmox Support Team Tue, 02 Dec 2025 01:38:32 +0100 proxmox-datacenter-manager (0.9.6) trixie; urgency=medium * server: fix-up jobstate directory permission mode to make state base dir use mode 0755 again. -- Proxmox Support Team Mon, 01 Dec 2025 05:22:42 +0100 proxmox-datacenter-manager (0.9.5) trixie; urgency=medium * views: add 'include-all' param; change semantics when there are no includes. * api-types: add 'layout' property to ViewConfig to prepare for UI integration. * api: implement CRUD endpoints for views. * api: require a minimum password length of 8 characters, following NIST recommendations. * api: pve remote: integrate read-only endpoints to get VM and CT configs. * api: firewall: add option, rules and status endpoints to prepare for initial support to show remote firewall config. * cli: restore functionality of the admin CLI tool. * api: add missing `returns` definitions to api macro tree-wide so this information shows up properly in a potential api viewer. * api: remote tasks: add fine-grained permission check on the remote level. * remote tasks: poll foreign, non-tracked active tasks with a low frequency to avoid them getting stuck. * remote tasks: make sure to update the task cache if there were errors when polling. * api: add endpoints to get remote apt system version and repository state. * views config: preparations for future glob/regex support. * server: add task log, auth log and access log rotation scheduled at midnight. -- Proxmox Support Team Mon, 01 Dec 2025 04:13:31 +0100 proxmox-datacenter-manager (0.9.4) trixie; urgency=medium * api: add new remote shell module with termproxy endpoints, which can be used to proxy a websocket based connection from PDM UI to the remote node. * api: pve: wire up remote shell support. * api: pbs: wire up node shell endpoints. * remote tasks: correctly filter out foreign running PBS tasks, avoiding suggesting that there seemingly is a steadily increasing amount of running remote tasks. * api: pve guests: fix regression with cross-remote migration * api: pve guest migrate: make passing storage and bridge mappings more robust. -- Proxmox Support Team Wed, 19 Nov 2025 21:32:12 +0100 proxmox-datacenter-manager (0.9.3) trixie; urgency=medium * server: use types indead of string for migration parameters. * fix #6901: api: add explicit permission check for SDN zones, vnets and controllers list endpoints. * api: subscription status: use correct resource ACL path in permission check. * api: use HTTP 403 Forbidden response code if user does have access to resource over the previously wrongly used 401 Unauthorized code. * api: add endpoints for querying the PDM node's status. * improve remote UPID handling, making it more type safe. * api, pbs-client: add bindings for task list, task status, task log and add PBS task API endpoints similar to the PVE ones. * remote updates, upgrade checks: fix old-version being now correctly being marked as optional. * add initial backend support for views, not yet exposed in the UI. * handle new unknown enum variants for better forward compatibility with newer remote project releases and features. -- Proxmox Support Team Thu, 13 Nov 2025 22:54:13 +0100 proxmox-datacenter-manager (0.9.2) trixie; urgency=medium * fix #6794: api: allow non-root users to view stats on dashboard. * api: resources: fix subscription level for remotes where all nodes have an unknown status. * fix #6901: api: add permission checks for PBS metrics endpoints. * fix #6901: PBS: fix permission check for showing overview by dropping the bogus `node` parameter from the ACL object path. * api: add support for adding OpenID Connect realms and using them for login. * api: add endpoints for managing API tokens. Please note that the PDM API is not yet stabilized. * PBS: request latest metrics when using hourly metrics time frame, just like we do it for PVE remotes. * resources: make search terms case insensitive. * update client dependencies to fix compatibility issue with some PBS 3 versions that wrongly signalled support for HttpOnly cookies. * metric collection: include PBS nodes in top-usage entities. * clean up ACL tree entries and api tokens for the removed user when deleting a user. * api daemon: implement cache for pending system updates of remotes and query the update status from all remotes every six hours in a background task. * api: add remote-type search category for resources to allow filtering for all resources from a specific remote type like PVE or PBS. * api: expose more status information and counters specific to PBS, including maintenance mode. * api: add property matching for resources and expose them for usage in the resource search system. * improve logging of failed remote API requests by including the full request path and also explicitly log unknown errors. -- Proxmox Support Team Mon, 27 Oct 2025 12:53:59 +0100 proxmox-datacenter-manager (0.9.1) trixie; urgency=medium * TTY login banner: fix typo. * server: api: add TLS probe endpoint for Proxmox Backup Server and implement endpoint to scan remote PBS instances to allow using the remote wizard to add PBS remotes. * add initial support to configure LDAP and AD authentication realms. * pbs-client: use `json` formatter for the PBS snapshot list streaming API. * api: add endpoint to list all namespaces from a datastore of a PBS remote. -- Proxmox Support Team Fri, 26 Sep 2025 19:41:41 +0200 proxmox-datacenter-manager (0.9.0) trixie; urgency=medium * bump version for BETA release based on Debian 13 Trixie. * self-signed cert: use correct product name for "Organization" field when creating a new certificate. -- Proxmox Support Team Thu, 11 Sep 2025 11:17:16 +0200 proxmox-datacenter-manager (0.3.2) trixie; urgency=medium * add initial pdmAtoB upgrade checker script. * cli/admin: add a versions command to show current package versions. * api: resources: implement remote match for template. * api: resources: improve behavior when searching for remotes to more easily allow to filter by resources of a specific remote, rather than limiting to the remote type. * api: resources: use root cause for errors that happen when fetching resources from a remote. -- Proxmox Support Team Wed, 10 Sep 2025 17:25:33 +0200 proxmox-datacenter-manager (0.3.1) trixie; urgency=medium * api: implement PVE storage usage metrics. * api: add nodes/storage endpoints for status and metrics. * api: add resource-type parameter when listing resources and include SDN Zones as resources. -- Proxmox Support Team Wed, 10 Sep 2025 12:08:54 +0200 proxmox-datacenter-manager (0.3.0) trixie; urgency=medium * update PDM apt sources to deb822 format and point to trixie. * server: api: resources: handle virtual guest template in status count. * metric collection: rework metric polling to run concurrently to improve total time required to poll all remotes, especially on high network latency and naturally also high remote counts. * metric collection: persist state after metric collection to allow reusing this after a daemon restart, e.g. due to a package update. * metric collection: collect overdue metrics on startup/timer change. * metric collection: record the remote response time and time required for a full collection run to provide better insights. * api: add endpoint to trigger metric collection. * api: remotes: trigger immediate metric collection for newly added nodes. * cli: add commands for metric-collection trigger and status. * metric collection: just skip missed timer ticks, no point in increasing polling frequency in such a case. * add ACL API by reusing common proxmox-access-control implementation. * remote task archive: handle missing journal file gracefully. * add api for getting available updates/changelogs for remote nodes. * add various (API) infrastructure for initial SDN integration. * api: resources: add more complex filter syntax that is loosely based on the Elasticsearch and GitHub query language. It basically works like: + 'sometext' : normal search term (OR'd with other terms) + '+sometext' : required search term + 'cat:text' : looks for 'text' in the category 'cat' * task cache: treat a limit of 0 as unbounded and return all tasks in that case. * api: remote tasks: add 'remote' filter option. * api: add remote-tasks statistics. -- Proxmox Support Team Thu, 04 Sep 2025 20:28:32 +0200 proxmox-datacenter-manager (0.2.2) trixie; urgency=medium * api server: make incoming connection handling more robust for some error cases. * qemu: enable connection tracking state transfer on VM remote live migration. Note that this is best effort as the network equipment and especially third party firewall in the target environment can still cause resets for open connections. * remote tasks: implement improved cache for remote tasks to split the single file cache into an 'active' and multiple 'archive' files. This improves lookup performance by around three orders of magnitude (from seconds to milliseconds) on benchmarks with 1.5 million tasks in the cache. Further the data structure was optimized for space, avoiding redundant information and older archived tasks will get compressed using zstd, which results in some significant reduction of storage space usage. This helps especially when there are many tasks, like when using many remotes with a big amount of guests. * remote tasks: add background task for polling task from all configured remotes every 10 minutes, use new task cache. Active tasks are tracked separately and their status is re-queried every 10 seconds. * api: add probe-tls endpoint to allow trust on first use (TOFU) prompts for when adding remotes. * api: acme cert order: fix constructing config with "default" account. -- Proxmox Support Team Wed, 27 Aug 2025 19:34:20 +0200 proxmox-datacenter-manager (0.2.1) trixie; urgency=medium * re-enable PBS integration again after fixing compilation issues in crates shared with UI. -- Proxmox Support Team Wed, 06 Aug 2025 20:14:35 +0200 proxmox-datacenter-manager (0.2.0) trixie; urgency=medium * initial upgrade for Debian 13 "Trixie" based release. -- Proxmox Support Team Tue, 05 Aug 2025 21:38:34 +0200 proxmox-datacenter-manager (0.1.11) bookworm; urgency=medium * privileged-api: create /var/cache/proxmox-datacenter-manager/ on startup * api: add caching 'remote-tasks' API endpoint * pve api: track new tasks created by PDM * api: fix remote upid tracking for qemu remote migration * api: add target-endpoint parameter to remote migrate api calls * api: pve: add remote cluster-status api call * client: add cluster status method * client: add target-endpoint parameter to remote migration methods * task cache: skip remotes with errors on fetch to avoid that a single unavailable cluster node can block the whole remote. * task cache: skip remotes with errors on fetch * task cache: add filter options and reverse task order * add optional web-url configuration for remotes to override the URL that the PDM web UI should use to open external resources. -- Proxmox Support Team Wed, 29 Jan 2025 09:21:27 +0100 proxmox-datacenter-manager (0.1.10) bookworm; urgency=medium * privileged api: do not clean-up socket unconditionally on start-up but only before freshly creating it. This should fix an issue where the unprivileged, main API dameon could not proxy to the privileged API daemon anymore after the latter was reloaded. Which, for example, can happen when installing package updates. * privileged api: output log message before signaling readiness to systemd. -- Proxmox Support Team Mon, 30 Dec 2024 15:43:10 +0100 proxmox-datacenter-manager (0.1.9) bookworm; urgency=medium * api: apt versions: fix version of running PDM daemon * api: apt versions: track more packages again * privileged api server: properly handle socket on reload * packaging: properly reload services after upgrade -- Proxmox Support Team Mon, 23 Dec 2024 16:32:26 +0100 proxmox-datacenter-manager (0.1.8) bookworm; urgency=medium * depend on libproxmox-acme-plugins package to ensure DNS challenge plugins are available. -- Proxmox Support Team Thu, 19 Dec 2024 18:56:31 +0100 proxmox-datacenter-manager (0.1.7) bookworm; urgency=medium * api: add more permission checks for dashboard calls so that those are usable for non-root users. the top entities are still root only for now. * remote migrate: fix handling of host:port as remote * remote migrate: don't log parameters * api: remotes: don't return token secret in response * server: api: add 'realms' add point for PVE -- Proxmox Support Team Thu, 19 Dec 2024 13:39:02 +0100 proxmox-datacenter-manager (0.1.6) bookworm; urgency=medium * explicitly depend on proxmox-termproxy. * ship GRUB_DISTRIBUTOR default-snippet for easier recognition on boot. -- Proxmox Support Team Wed, 18 Dec 2024 23:38:43 +0100 proxmox-datacenter-manager (0.1.5) bookworm; urgency=medium * api: log listening address and port used * api: use always use base user when trying to create a token * packages: recommend UI package from API and vice versa * add simple daily-update executable with systemd service & timer * fix installing apt sources file for pdm-test repo * api: add qemu migrate preconditions api call * automatically poll all remotes and fill metrics cache every 15 minutes. -- Proxmox Support Team Wed, 18 Dec 2024 15:55:21 +0100 proxmox-datacenter-manager (0.1.4) bookworm; urgency=medium * api: add banner service to write our /etc/issue -- Proxmox Support Team Wed, 18 Dec 2024 10:40:24 +0100 proxmox-datacenter-manager (0.1.3) bookworm; urgency=medium * api: pve: various permission check improvements * server: remove debug logging from calculating the top entities * server: api: include more 'top entities' metrics * server: api: top entities: make timeframe configurable * ship sources.list file with pdm-test repo configured -- Proxmox Support Team Wed, 18 Dec 2024 09:47:27 +0100 proxmox-datacenter-manager (0.1.2) bookworm; urgency=medium * admin cli: set up server context on startup * server: add docgen binary * server: api: add pve node status api call * record various missing package dependencies -- Proxmox Support Team Mon, 16 Dec 2024 21:26:08 +0100 proxmox-datacenter-manager (0.1.1) bookworm; urgency=medium * bump version for testing packaging with current development state. -- Proxmox Support Team Sun, 15 Dec 2024 21:10:19 +0100 proxmox-datacenter-manager (0.1.0) bookworm; urgency=medium * initial release -- Proxmox Support Team Thu, 26 Oct 2023 18:44:35 +0200