1 pve-ha-manager (3.5.1) bullseye; urgency=medium
3 * manager: update crs scheduling mode once per round to avoid the need for a
4 restart of the currently active manager.
6 * api: status: add CRS info to manager if not set to default
8 -- Proxmox Support Team <support@proxmox.com> Sat, 19 Nov 2022 15:51:11 +0100
10 pve-ha-manager (3.5.0) bullseye; urgency=medium
12 * env: datacenter config: include crs (cluster-resource-scheduling) setting
14 * manager: use static resource scheduler when configured
16 * manager: avoid scoring nodes if maintenance fallback node is valid
18 * manager: avoid scoring nodes when not trying next and current node is
21 * usage: static: use service count on nodes as a fallback
23 -- Proxmox Support Team <support@proxmox.com> Fri, 18 Nov 2022 15:02:55 +0100
25 pve-ha-manager (3.4.0) bullseye; urgency=medium
27 * switch to native version formatting
29 * fix accounting of online services when moving services due to their source
30 node going gracefully nonoperational (maintenance mode). This ensures a
31 better balance of services on the cluster after such an operation.
33 -- Proxmox Support Team <support@proxmox.com> Fri, 22 Jul 2022 09:21:20 +0200
35 pve-ha-manager (3.3-4) bullseye; urgency=medium
37 * lrm: fix getting stuck on restart due to finished worker state not
40 -- Proxmox Support Team <support@proxmox.com> Wed, 27 Apr 2022 14:01:55 +0200
42 pve-ha-manager (3.3-3) bullseye; urgency=medium
44 * lrm: avoid possible job starvation on huge workloads
46 * lrm: increase run_worker loop-time for doing actually work to 80%
49 -- Proxmox Support Team <support@proxmox.com> Thu, 20 Jan 2022 18:05:33 +0100
51 pve-ha-manager (3.3-2) bullseye; urgency=medium
53 * fix #3826: fix restarting LRM/CRM when triggered by package management
54 system due to other updates
56 * lrm: also check CRM node-status for determining if there's a fence-request
57 and avoid starting up in that case to ensure that the current manager can
58 get our lock and do a clean fence -> unknown -> online FSM transition.
59 This avoids a problematic edge case where an admin manually removed all
60 services of a to-be-fenced node, and re-added them again before the
61 manager could actually get that nodes LRM lock.
63 * manage: handle edge case where a node gets seemingly stuck in 'fence'
64 state if all its services got manually removed by an admin before the
65 fence transition could be finished. While the LRM could come up again in
66 previous versions (it won't now, see above point) and start/stop services
67 got executed, the node was seen as unavailable for all recovery,
68 relocation and migrate actions.
70 -- Proxmox Support Team <support@proxmox.com> Wed, 19 Jan 2022 14:30:15 +0100
72 pve-ha-manager (3.3-1) bullseye; urgency=medium
74 * LRM: release lock and close watchdog if no service configured for >10min
76 * manager: make recovery actual state in finite state machine, showing a
77 clear transition from fence -> reocvery.
79 * fix #3415: never switch in error state on recovery, try to find a new node
80 harder. This improves using the HA manager for services with local
81 resources (e.g., local storage) to ensure it always gets started, which is
82 an OK use-case as long as the service is restricted to a group with only
83 that node. Previously failure of that node would have a high possibility
84 of the service going into the errors state, as no new node can be found.
85 Now it will retry finding a new node, and if one of the restricted set,
86 e.g., the node it was previous on, comes back up, it will start again
89 * recovery: allow disabling a in-recovery service manually
91 -- Proxmox Support Team <support@proxmox.com> Fri, 02 Jul 2021 20:03:29 +0200
93 pve-ha-manager (3.2-2) bullseye; urgency=medium
95 * fix systemd service restart behavior on package upgrade with Debian
98 -- Proxmox Support Team <support@proxmox.com> Mon, 24 May 2021 11:38:42 +0200
100 pve-ha-manager (3.2-1) bullseye; urgency=medium
102 * Re-build for Debian Bullseye / PVE 7
104 -- Proxmox Support Team <support@proxmox.com> Wed, 12 May 2021 20:55:53 +0200
106 pve-ha-manager (3.1-1) pve; urgency=medium
108 * allow 'with-local-disks' migration for replicated guests
110 -- Proxmox Support Team <support@proxmox.com> Mon, 31 Aug 2020 10:52:23 +0200
112 pve-ha-manager (3.0-9) pve; urgency=medium
114 * factor out service configured/delete helpers
116 * typo and grammar fixes
118 -- Proxmox Support Team <support@proxmox.com> Thu, 12 Mar 2020 13:17:36 +0100
120 pve-ha-manager (3.0-8) pve; urgency=medium
122 * bump LRM stop wait time to an hour
124 * do not mark maintenaned nodes as unkown
126 * api/status: extra handling of maintenance mode
128 -- Proxmox Support Team <support@proxmox.com> Mon, 02 Dec 2019 10:33:03 +0100
130 pve-ha-manager (3.0-6) pve; urgency=medium
132 * add 'migrate' node shutdown policy
134 * do simple fallback if node comes back online from maintenance
136 * account service to both, source and target during migration
138 * add 'After' ordering for SSH and pveproxy to LRM service, ensuring the node
139 stays accessible until HA services got moved or shutdown, depending on
142 -- Proxmox Support Team <support@proxmox.com> Tue, 26 Nov 2019 18:03:26 +0100
144 pve-ha-manager (3.0-5) pve; urgency=medium
146 * fix #1339: remove more locks from services IF the node got fenced
148 * adapt to qemu-server code refactoring
150 -- Proxmox Support Team <support@proxmox.com> Wed, 20 Nov 2019 20:12:49 +0100
152 pve-ha-manager (3.0-4) pve; urgency=medium
154 * use PVE::DataCenterConfig from new split-out cluster library package
156 -- Proxmox Support Team <support@proxmox.com> Mon, 18 Nov 2019 12:16:29 +0100
158 pve-ha-manager (3.0-3) pve; urgency=medium
160 * fix #1919, #1920: improve handling zombie (without node) services
162 * fix # 2241: VM resource: allow migration with local device, when not running
164 * HA status: render removal transition of service as 'deleting'
166 * fix #1140: add crm command 'stop', which allows to request immediate
167 service hard-stops if a timeout of zero (0) is passed
169 -- Proxmox Support Team <support@proxmox.com> Mon, 11 Nov 2019 17:04:35 +0100
171 pve-ha-manager (3.0-2) pve; urgency=medium
173 * services: update PIDFile to point directly to /run
175 * fix #2234: fix typo in service description
177 * Add missing Dependencies to pve-ha-simulator
179 -- Proxmox Support Team <support@proxmox.com> Thu, 11 Jul 2019 19:26:03 +0200
181 pve-ha-manager (3.0-1) pve; urgency=medium
183 * handle the case where a node gets fully purged
185 * Re-build for Debian Buster / PVE 6
187 -- Proxmox Support Team <support@proxmox.com> Wed, 22 May 2019 19:11:59 +0200
189 pve-ha-manager (2.0-9) unstable; urgency=medium
191 * get_ha_settings: cope with (temporarily) unavailable pmxcfs
193 * lrm: exit on restart and agent lock lost for > 90s
195 * service data: only set failed_nodes key if needed
197 -- Proxmox Support Team <support@proxmox.com> Thu, 04 Apr 2019 16:27:32 +0200
199 pve-ha-manager (2.0-8) unstable; urgency=medium
201 * address an issue in dpkg 1.18 with wrong trigger cycle detections if cyclic
202 dependencies are involed
204 -- Proxmox Support Team <support@proxmox.com> Wed, 06 Mar 2019 07:49:58 +0100
206 pve-ha-manager (2.0-7) unstable; urgency=medium
208 * fix #1842: do not pass forceStop to CT shutdown
210 * fix #1602: allow one to delete ignored services over API
212 * fix #1891: Add zsh command completion for ha-manager CLI tools
214 * fix #1794: VM resource: catch qmp command exceptions
216 * show sent emails in regression tests
218 -- Proxmox Support Team <support@proxmox.com> Mon, 04 Mar 2019 10:37:25 +0100
220 pve-ha-manager (2.0-6) unstable; urgency=medium
222 * fix #1378: allow one to specify a service shutdown policy
224 * remove some unused external dependencies from the standalone simulator
227 * document api result for ha resources
229 -- Proxmox Support Team <support@proxmox.com> Mon, 07 Jan 2019 12:59:27 +0100
231 pve-ha-manager (2.0-5) unstable; urgency=medium
233 * skip CRM and LRM work if last cfs update failed
235 * regression test system: allow to simulate cluster fs failures
237 * postinst: drop transitional cleanup for systemd watchdog mux socket
239 -- Proxmox Support Team <support@proxmox.com> Wed, 07 Feb 2018 11:00:12 +0100
241 pve-ha-manager (2.0-4) unstable; urgency=medium
243 * address timing issues happening when pve-cluster.service is being restarted
245 -- Proxmox Support Team <support@proxmox.com> Thu, 09 Nov 2017 11:46:50 +0100
247 pve-ha-manager (2.0-3) unstable; urgency=medium
249 * add ignore state for resources
251 * lrm/crm service: restart on API changes
253 * lrm.service: do not timeout on stop
255 * fix #1347: let postfix fill in FQDN in fence mails
257 * fix #1073: do not count backup-suspended VMs as running
259 -- Proxmox Support Team <support@proxmox.com> Fri, 13 Oct 2017 11:10:51 +0200
261 pve-ha-manager (2.0-2) unstable; urgency=medium
263 * explicitly sync journal when disabling watchdog updates
265 * always queue service stop if node shuts down
267 * Fix shutdown order of HA and storage services
269 * Resource/API: abort early if resource in error state
271 -- Proxmox Support Team <support@proxmox.com> Wed, 14 Jun 2017 07:49:59 +0200
273 pve-ha-manager (2.0-1) unstable; urgency=medium
275 * rebuild for PVE 5.0 / Debian Stretch
277 -- Proxmox Support Team <support@proxmox.com> Mon, 13 Mar 2017 11:31:53 +0100
279 pve-ha-manager (1.0-40) unstable; urgency=medium
281 * ha-simulator: allow adding service on runtime
283 * ha-simulator: allow deleting service via GUI
285 * ha-simulator: allow new service request states over gui
287 * ha-simulator: use JSON instead of Dumper for manager status view
289 -- Proxmox Support Team <support@proxmox.com> Tue, 24 Jan 2017 10:03:07 +0100
291 pve-ha-manager (1.0-39) unstable; urgency=medium
293 * add setup_environment hook to CLIHandler class
295 * ha-simulator: fix typo s/Mode/Node/
297 * is_node_shutdown: check for correct systemd targets
299 * Simulator: fix scrolling to end of cluster log view
301 * Simulator: do not use cursor position to insert log
303 -- Proxmox Support Team <support@proxmox.com> Thu, 12 Jan 2017 13:15:08 +0100
305 pve-ha-manager (1.0-38) unstable; urgency=medium
309 -- Proxmox Support Team <support@proxmox.com> Wed, 23 Nov 2016 11:46:21 +0100
311 pve-ha-manager (1.0-37) unstable; urgency=medium
313 * HA::Status: provide better/faster feedback
315 * Manager.pm: store flag to indicate successful start
317 * ha status: include common service attributes
319 * Groups.pm: add verbose_description for 'restricted'
321 * Resources.pm: use verbose_description for state
323 * pve-ha-group-node-list: add verbose_description
325 * ha-manager: remove 'enabled' and 'disabled' commands
327 * rename request state 'enabled' to 'started'
329 * get_pve_lock: correctly send a lock update request
331 -- Proxmox Support Team <support@proxmox.com> Tue, 22 Nov 2016 17:04:57 +0100
333 pve-ha-manager (1.0-36) unstable; urgency=medium
335 * Resources: implement 'stopped' state
337 * ha-manager: remove obsolet pod content
339 * Fix #1189: correct spelling in fence mail
341 * API/Status: avoid using HA Environment
343 * factor out resource config check and default set code
345 -- Proxmox Support Team <support@proxmox.com> Tue, 15 Nov 2016 16:42:07 +0100
347 pve-ha-manager (1.0-35) unstable; urgency=medium
349 * change service state to error if no recovery node is available
351 * cleanup backup & mounted locks after recovery (fixes #1100)
353 * add possibility to simulate locks from services
355 * don't run regression test when building the simulator package
357 -- Proxmox Support Team <support@proxmox.com> Thu, 15 Sep 2016 13:23:00 +0200
359 pve-ha-manager (1.0-34) unstable; urgency=medium
361 * fix race condition on slow resource commands in started state
363 -- Proxmox Support Team <support@proxmox.com> Mon, 12 Sep 2016 13:07:05 +0200
365 pve-ha-manager (1.0-33) unstable; urgency=medium
367 * relocate policy: try to avoid already failed nodes
369 * allow empty json status files
371 * more regression tests
373 -- Proxmox Support Team <support@proxmox.com> Fri, 22 Jul 2016 12:16:48 +0200
375 pve-ha-manager (1.0-32) unstable; urgency=medium
377 * use correct verify function for ha-group-node-list
379 * send email on fence failure and success
381 -- Proxmox Support Team <support@proxmox.com> Wed, 15 Jun 2016 17:01:12 +0200
383 pve-ha-manager (1.0-31) unstable; urgency=medium
385 * selcet_service_node: include all online nodes in default group
387 * LRM: do not count erroneous service as active
389 * fix relocate/restart trial count leak on service deletion
391 -- Proxmox Support Team <support@proxmox.com> Fri, 06 May 2016 08:26:10 +0200
393 pve-ha-manager (1.0-30) unstable; urgency=medium
395 * Env: allow debug logging
397 -- Proxmox Support Team <support@proxmox.com> Fri, 29 Apr 2016 16:50:34 +0200
399 pve-ha-manager (1.0-29) unstable; urgency=medium
401 * Resources: deny setting nonexistent group
403 -- Proxmox Support Team <support@proxmox.com> Wed, 20 Apr 2016 18:22:28 +0200
405 pve-ha-manager (1.0-28) unstable; urgency=medium
407 * Config: add get_service_status method
409 -- Proxmox Support Team <support@proxmox.com> Tue, 19 Apr 2016 08:41:22 +0200
411 pve-ha-manager (1.0-27) unstable; urgency=medium
413 * use pve-doc-generator to generate man pages
415 -- Proxmox Support Team <support@proxmox.com> Fri, 08 Apr 2016 08:25:07 +0200
417 pve-ha-manager (1.0-26) unstable; urgency=medium
419 * status: show added but not yet active services
421 * status: mark CRM as idle if no service is configured
423 -- Proxmox Support Team <support@proxmox.com> Tue, 15 Mar 2016 12:49:18 +0100
425 pve-ha-manager (1.0-25) unstable; urgency=medium
427 * Use config_file from PVE::QemuConfig
429 -- Proxmox Support Team <support@proxmox.com> Tue, 08 Mar 2016 11:50:49 +0100
431 pve-ha-manager (1.0-24) unstable; urgency=medium
433 * simulator: install all virtual resources
435 -- Proxmox Support Team <support@proxmox.com> Wed, 02 Mar 2016 10:30:40 +0100
437 pve-ha-manager (1.0-23) unstable; urgency=medium
439 * fix infinite started <=> migrate cycle
441 * exec_resource_agent: process error state early
443 * avoid out of sync command execution in LRM
445 * do not pass ETRY_AGAIN back to the CRM
447 -- Proxmox Support Team <support@proxmox.com> Wed, 24 Feb 2016 12:15:21 +0100
449 pve-ha-manager (1.0-22) unstable; urgency=medium
451 * fix 'change_service_location' misuse and recovery from fencing
453 * add VirtFail resource and use it in new regression tests
455 * improve relocation policy code in manager and LRM
457 * improve verbosity of API status call
459 -- Proxmox Support Team <support@proxmox.com> Mon, 15 Feb 2016 10:57:44 +0100
461 pve-ha-manager (1.0-21) unstable; urgency=medium
463 * Fix postinstall script not removing watchdog-mux.socket
465 -- Proxmox Support Team <support@proxmox.com> Thu, 04 Feb 2016 18:23:47 +0100
467 pve-ha-manager (1.0-20) unstable; urgency=medium
469 * LRM: do not release lock on shutdown errors
471 * Split up resources and move them to own sub folder
473 * Add virtual resources for tests and simulation
475 * add after_fork method to HA environment and use it in LRM
477 -- Proxmox Support Team <support@proxmox.com> Wed, 27 Jan 2016 17:05:23 +0100
479 pve-ha-manager (1.0-19) unstable; urgency=medium
481 * remove 'running' from migrate/relocate log message
483 * LRM: release agent lock on graceful shutdown
485 * LRM: release agent lock also on restart
487 * CRM: release lock on shutdown request
489 * TestHardware: correct shutdown/reboot behaviour of CRM and LRM
491 * resource agents: fix relocate
493 -- Proxmox Support Team <support@proxmox.com> Mon, 18 Jan 2016 12:41:08 +0100
495 pve-ha-manager (1.0-18) unstable; urgency=medium
497 * pve-ha-lrm.service: depend on lxc.service
499 * output watchdog module name if it gets loaded
501 * remove watchdog-mux.socket
503 -- Proxmox Support Team <support@proxmox.com> Tue, 12 Jan 2016 12:27:49 +0100
505 pve-ha-manager (1.0-17) unstable; urgency=medium
507 * Resources.pm: use PVE::API2::LXC
509 -- Proxmox Support Team <support@proxmox.com> Mon, 11 Jan 2016 12:25:38 +0100
511 pve-ha-manager (1.0-16) unstable; urgency=medium
513 * check_active_workers: fix typo /uuid/uid/
515 -- Proxmox Support Team <support@proxmox.com> Mon, 21 Dec 2015 10:21:30 +0100
517 pve-ha-manager (1.0-15) unstable; urgency=medium
519 * stop all resources on node shutdown (instead of freeze)
521 -- Proxmox Support Team <support@proxmox.com> Wed, 16 Dec 2015 10:33:30 +0100
523 pve-ha-manager (1.0-14) unstable; urgency=medium
525 * allow to configure watchdog module in /etc/default/pve-ha-manager
527 -- Proxmox Support Team <support@proxmox.com> Thu, 03 Dec 2015 11:09:47 +0100
529 pve-ha-manager (1.0-13) unstable; urgency=medium
531 * HA API: Fix permissions
533 -- Proxmox Support Team <support@proxmox.com> Fri, 30 Oct 2015 11:16:50 +0100
535 pve-ha-manager (1.0-12) unstable; urgency=medium
537 * Adding constants to gain more readability
539 * exec_resource_agent: return valid exit code instead of die's
543 -- Proxmox Support Team <support@proxmox.com> Thu, 29 Oct 2015 10:21:49 +0100
545 pve-ha-manager (1.0-11) unstable; urgency=medium
547 * add workaround for bug #775
549 -- Proxmox Support Team <support@proxmox.com> Wed, 21 Oct 2015 08:58:41 +0200
551 pve-ha-manager (1.0-10) unstable; urgency=medium
553 * better resource status check on addition and update
555 -- Proxmox Support Team <support@proxmox.com> Mon, 12 Oct 2015 18:26:24 +0200
557 pve-ha-manager (1.0-9) unstable; urgency=medium
559 * delete node from CRM status when deleted from cluster
561 -- Proxmox Support Team <support@proxmox.com> Tue, 29 Sep 2015 07:35:30 +0200
563 pve-ha-manager (1.0-8) unstable; urgency=medium
565 * Use new lock domain sub instead of storage lock
567 -- Proxmox Support Team <support@proxmox.com> Sat, 26 Sep 2015 10:36:09 +0200
569 pve-ha-manager (1.0-7) unstable; urgency=medium
571 * enhance ha-managers group commands
573 * vm_is_ha_managed: allow check on service state
577 -- Proxmox Support Team <support@proxmox.com> Mon, 21 Sep 2015 12:17:41 +0200
579 pve-ha-manager (1.0-6) unstable; urgency=medium
581 * implement bash completion for ha-manager
583 * implement recovery policy for services
585 * simulator: fix random output of manager status
587 -- Proxmox Support Team <support@proxmox.com> Wed, 16 Sep 2015 12:06:12 +0200
589 pve-ha-manager (1.0-5) unstable; urgency=medium
591 * Replacing hardcoded qemu commands with plugin calls
593 * improve error state behaviour
595 -- Proxmox Support Team <support@proxmox.com> Tue, 08 Sep 2015 08:45:36 +0200
597 pve-ha-manager (1.0-4) unstable; urgency=medium
599 * groups: encode nodes as hash (internally)
601 * add trigger for pve-api-updates
603 -- Proxmox Support Team <support@proxmox.com> Tue, 16 Jun 2015 09:59:03 +0200
605 pve-ha-manager (1.0-3) unstable; urgency=medium
607 * CRM: do not start if there is no resource.cfg file to avoid warnings
609 -- Proxmox Support Team <support@proxmox.com> Tue, 09 Jun 2015 14:35:09 +0200
611 pve-ha-manager (1.0-2) unstable; urgency=medium
613 * use Wants instead of Requires inside systemd service definitions
615 -- Proxmox Support Team <support@proxmox.com> Tue, 09 Jun 2015 09:33:24 +0200
617 pve-ha-manager (1.0-1) unstable; urgency=medium
619 * enable/start crm and lrm services by default
621 -- Proxmox Support Team <support@proxmox.com> Fri, 05 Jun 2015 10:03:53 +0200
623 pve-ha-manager (0.9-3) unstable; urgency=medium
625 * regression test improvements
627 -- Proxmox Support Team <support@proxmox.com> Fri, 10 Apr 2015 06:53:23 +0200
629 pve-ha-manager (0.9-2) unstable; urgency=medium
631 * issue warning if ha group does not exist
633 -- Proxmox Support Team <support@proxmox.com> Tue, 07 Apr 2015 09:52:07 +0200
635 pve-ha-manager (0.9-1) unstable; urgency=medium
637 * rename vm resource prefix: pvevm: => vm:
639 * add API to query ha status
641 * allow to use simply VMIDs as resource id
645 -- Proxmox Support Team <support@proxmox.com> Fri, 03 Apr 2015 06:18:05 +0200
647 pve-ha-manager (0.8-2) unstable; urgency=medium
649 * lrm: reduce TimeoutStopSec to 95
651 * lrm: set systemd killmode to 'process'
653 -- Proxmox Support Team <support@proxmox.com> Thu, 02 Apr 2015 08:48:24 +0200
655 pve-ha-manager (0.8-1) unstable; urgency=medium
657 * currecrtly send cfs lock update request
659 -- Proxmox Support Team <support@proxmox.com> Thu, 02 Apr 2015 08:18:00 +0200
661 pve-ha-manager (0.7-1) unstable; urgency=medium
663 * create /etc/pve/ha automatically
665 * use correct package for lock_ha_config
667 * fix ha-manager status when ha is unconfigured
669 * do not unlink watchdog socket when started via systemd
673 -- Proxmox Support Team <support@proxmox.com> Wed, 01 Apr 2015 11:05:08 +0200
675 pve-ha-manager (0.6-1) unstable; urgency=medium
677 * move configuration handling into PVE::HA::Config
679 * ha-manager status: add --verbose flag
681 * depend on qemu-server
683 -- Proxmox Support Team <support@proxmox.com> Fri, 27 Mar 2015 12:28:50 +0100
685 pve-ha-manager (0.5-1) unstable; urgency=medium
687 * implement service migration
689 * fix service dependencies (allow restart, reboot)
691 * freeze services during reboot/restart
693 -- Proxmox Support Team <support@proxmox.com> Thu, 26 Mar 2015 13:22:58 +0100
695 pve-ha-manager (0.4-1) unstable; urgency=medium
697 * increase fence_delay to 60 seconds
699 * fix regression test environment
701 * fix failover after master crash with pending fence action
703 -- Proxmox Support Team <support@proxmox.com> Wed, 25 Mar 2015 13:59:28 +0100
705 pve-ha-manager (0.3-1) unstable; urgency=medium
707 * really activate softdog
709 * correctly count active services
711 * implement fence_delay to avoid immediate fencing
713 * pve-ha-simulator: reset watchdog with poweroff
715 * pve-ha-simulator: use option nofailback for default groups
717 -- Proxmox Support Team <support@proxmox.com> Mon, 16 Mar 2015 13:03:23 +0100
719 pve-ha-manager (0.2-1) unstable; urgency=medium
721 * add ha-manager command line tool
723 * start implementing resources and groups API
725 -- Proxmox Support Team <support@proxmox.com> Fri, 13 Mar 2015 09:26:12 +0100
727 pve-ha-manager (0.1-1) unstable; urgency=low
731 -- Proxmox Support Team <support@proxmox.com> Wed, 18 Feb 2015 11:30:21 +0100