* [PATCH pve-manager v5 01/10] fix #6144: api: nodes: add endpoint to set maintenance mode
2026-11-19 9:59 [PATCH-SERIES docs/ha-manager/manager v5 00/10] fix #6144: add api and ui for node maintenance mode Dominik Rusovac
@ 2026-11-19 9:59 ` Dominik Rusovac
2026-11-19 9:59 ` [PATCH pve-manager v5 02/10] fix #6144: ui: ha: status: allow toggling " Dominik Rusovac
` (9 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: Dominik Rusovac @ 2026-11-19 9:59 UTC (permalink / raw)
To: pve-devel; +Cc: Thomas Skinner
node maintenance mode can currently only be requested with the HA
manager's CLI, which can't be used from the web UI
add an endpoint to enable/disable maintenance mode for a node, which
lives below the node path rather than the HA API, so that the
maintenance concept can be extended to more than HA managed resources
later on
for now, it runs the HA manager's CLI command to request the HA node
maintenance mode. the CLI rejects requests that the CRM would ignore,
e.g. for nodes unknown to HA, while HA is disarmed or if the node is not
online, and its error message is forwarded
Originally-by: Thomas Skinner <thomas@atskinner.net>
Signed-off-by: Dominik Rusovac <d.rusovac@proxmox.com>
---
Notes:
changes since v4:
* reword commit subject and body
* reorder name, path, method, ... etc.
* reword 'maintenance-mode' description
* run the HA manager's CLI command instead of reading the HA
manager status and queuing the CRM command directly, as
suggested in the v4 review, and forward its error message
PVE/API2/Nodes.pm | 45 +++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 45 insertions(+)
diff --git a/PVE/API2/Nodes.pm b/PVE/API2/Nodes.pm
index 07b3ce1f..117daa45 100644
--- a/PVE/API2/Nodes.pm
+++ b/PVE/API2/Nodes.pm
@@ -249,6 +249,7 @@ __PACKAGE__->register_method({
{ name => 'hosts' },
{ name => 'journal' },
{ name => 'lxc' },
+ { name => 'maintenance' },
{ name => 'migrateall' },
{ name => 'netstat' },
{ name => 'network' },
@@ -2908,6 +2909,50 @@ __PACKAGE__->register_method({
},
});
+__PACKAGE__->register_method({
+ name => 'maintenance_set',
+ path => 'maintenance',
+ method => 'PUT',
+ description => "Enable or disable maintenance mode for the node. Currently, this only"
+ . " requests the HA node maintenance mode.",
+ protected => 1,
+ permissions => {
+ check => ['perm', '/nodes/{node}', ['Sys.Modify']],
+ },
+ parameters => {
+ additionalProperties => 0,
+ properties => {
+ node => get_standard_option('pve-node'),
+ 'maintenance-mode' => {
+ description => "Specifies whether the maintenance mode for the specified"
+ . " node should be enabled or disabled.",
+ type => 'string',
+ enum => ['enable', 'disable'],
+ },
+ },
+ },
+ returns => { type => 'null' },
+ code => sub {
+ my ($param) = @_;
+
+ my $node = $param->{node};
+ my $mode = $param->{'maintenance-mode'};
+
+ my $err = '';
+ eval {
+ PVE::Tools::run_command(
+ [
+ 'ha-manager', 'crm-command', 'node-maintenance', $mode, $node,
+ ],
+ errfunc => sub { $err .= "$_[0]\n" },
+ );
+ };
+ die($err || $@) if $@;
+
+ return undef;
+ },
+});
+
# bash completion helper
sub complete_templet_repo {
--
2.47.3
^ permalink raw reply related [flat|nested] 12+ messages in thread* [PATCH pve-manager v5 02/10] fix #6144: ui: ha: status: allow toggling maintenance mode
2026-11-19 9:59 [PATCH-SERIES docs/ha-manager/manager v5 00/10] fix #6144: add api and ui for node maintenance mode Dominik Rusovac
2026-11-19 9:59 ` [PATCH pve-manager v5 01/10] fix #6144: api: nodes: add endpoint to set " Dominik Rusovac
@ 2026-11-19 9:59 ` Dominik Rusovac
2026-11-19 9:59 ` [PATCH pve-manager v5 03/10] ui: ha: status: show spinner while maintenance mode request is queued Dominik Rusovac
` (8 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: Dominik Rusovac @ 2026-11-19 9:59 UTC (permalink / raw)
To: pve-devel; +Cc: Thomas Skinner
maintenance mode could only be enabled or disabled through the HA
manager's CLI
add a 'Node Maintenance' menu to the HA status panel, which requests
enabling/disabling maintenance mode for the node of the selected lrm
entry
enabling is not offered while HA is disarmed, as the CRM ignores such
requests in that state. menu is disabled for users lacking the
Sys.Modify privilege on nodes
the LRM also enters maintenance mode on shutdown with the 'migrate'
shutdown policy, which a disable request would not undo
'fa fa-check' icon was chosen for 'Disable' option to be consistent
with which icon a node gets when coming back from maintenance mode
Originally-by: Thomas Skinner <thomas@atskinner.net>
Signed-off-by: Dominik Rusovac <d.rusovac@proxmox.com>
---
Notes:
changes since v4:
* rename maintenanceNode to selectedLrmNode and comment which node
the other maintenance field of the view model refers to
* throw on unknown maintenance modes instead of defaulting to enable
* split out the spinner into a separate patch
* reword commit body accordingly
www/manager6/ha/Status.js | 98 +++++++++++++++++++++++++++++++++++
www/manager6/ha/StatusView.js | 4 ++
2 files changed, 102 insertions(+)
diff --git a/www/manager6/ha/Status.js b/www/manager6/ha/Status.js
index c01bd737..42e9aefb 100644
--- a/www/manager6/ha/Status.js
+++ b/www/manager6/ha/Status.js
@@ -11,6 +11,15 @@ Ext.define('PVE.ha.Status', {
viewModel: {
data: {
haDisarmed: false,
+ selectedLrmNode: null,
+ // maintenance state of the selected LRM node
+ maintenanceRequested: false,
+ },
+ formulas: {
+ noMaintenanceAction: (get) =>
+ !get('selectedLrmNode') || !Ext.state.Manager.get('GuiCap').nodes['Sys.Modify'],
+ // the CRM ignores requests to enable maintenance mode while HA is disarmed
+ noMaintenanceEnable: (get) => get('maintenanceRequested') || get('haDisarmed'),
},
},
@@ -52,6 +61,65 @@ Ext.define('PVE.ha.Status', {
);
},
+ updateMaintenanceState: function () {
+ let me = this;
+ let vm = me.getViewModel();
+ let sv = me.getView().down('pveHAStatusView');
+ let rec = sv.getSelection()[0];
+ let isLrm = rec?.get('type') === 'lrm';
+
+ vm.set('selectedLrmNode', isLrm ? rec.get('node') : null);
+ vm.set('maintenanceRequested', isLrm && rec.get('maintenance-requested'));
+ },
+
+ handleMaintenanceButton: function (menuItem) {
+ let me = this;
+ let node = me.getViewModel().get('selectedLrmNode');
+ if (!node) {
+ return;
+ }
+
+ let msg;
+ if (menuItem.maintenanceMode === 'disable') {
+ msg = Ext.String.format(
+ gettext("Are you sure you want to disable maintenance mode for node '{0}'?"),
+ node,
+ );
+ if (me.getViewModel().get('haDisarmed')) {
+ msg +=
+ '<br><br>' +
+ Ext.String.format(
+ gettext(
+ "HA resources will be migrated back to node '{0}' once HA is re-armed.",
+ ),
+ node,
+ );
+ }
+ } else if (menuItem.maintenanceMode === 'enable') {
+ msg = Ext.String.format(
+ gettext("Are you sure you want to enable maintenance mode for node '{0}'?"),
+ node,
+ );
+ msg +=
+ '<br><br>' +
+ gettext('HA resources on the node will be migrated to other nodes.');
+ } else {
+ throw `unknown maintenance mode '${menuItem.maintenanceMode}'`;
+ }
+
+ Ext.Msg.confirm(gettext('Confirm'), msg, function (btn) {
+ if (btn !== 'yes') {
+ return;
+ }
+ Proxmox.Utils.API2Request({
+ url: `/nodes/${node}/maintenance`,
+ params: { 'maintenance-mode': menuItem.maintenanceMode },
+ method: 'PUT',
+ failure: (response) => Ext.Msg.alert(gettext('Error'), response.htmlStatus),
+ });
+ });
+ },
+
openCRSOptions: function () {
Ext.create('PVE.form.CRSOptions', {
autoShow: true,
@@ -121,6 +189,34 @@ Ext.define('PVE.ha.Status', {
},
],
},
+ '-',
+ {
+ text: gettext('Node Maintenance'),
+ iconCls: 'fa fa-wrench',
+ bind: {
+ disabled: '{noMaintenanceAction}',
+ },
+ menu: [
+ {
+ text: gettext('Enable'),
+ iconCls: 'fa fa-wrench',
+ maintenanceMode: 'enable',
+ bind: {
+ disabled: '{noMaintenanceEnable}',
+ },
+ handler: 'handleMaintenanceButton',
+ },
+ {
+ text: gettext('Disable'),
+ iconCls: 'fa fa-check',
+ maintenanceMode: 'disable',
+ bind: {
+ disabled: '{!maintenanceRequested}',
+ },
+ handler: 'handleMaintenanceButton',
+ },
+ ],
+ },
'->',
{
text: gettext('CRS Settings'),
@@ -153,11 +249,13 @@ Ext.define('PVE.ha.Status', {
collapsible: true,
padding: '0 0 20 0',
listeners: {
+ selectionchange: () => me.getController().updateMaintenanceState(),
hastatuschange: function (isDisarmed) {
let vm = me.getViewModel();
let sv = me.query('pveHAStatusView')[0];
vm.set('haDisarmed', isDisarmed);
+ me.getController().updateMaintenanceState();
if (sv.isDisarmedPendingState === null) {
return;
diff --git a/www/manager6/ha/StatusView.js b/www/manager6/ha/StatusView.js
index f260bc0d..d08534b4 100644
--- a/www/manager6/ha/StatusView.js
+++ b/www/manager6/ha/StatusView.js
@@ -132,6 +132,10 @@ Ext.define(
},
},
'armed-state',
+ {
+ name: 'maintenance-requested',
+ type: 'boolean',
+ },
],
idProperty: 'id',
});
--
2.47.3
^ permalink raw reply related [flat|nested] 12+ messages in thread* [PATCH pve-manager v5 03/10] ui: ha: status: show spinner while maintenance mode request is queued
2026-11-19 9:59 [PATCH-SERIES docs/ha-manager/manager v5 00/10] fix #6144: add api and ui for node maintenance mode Dominik Rusovac
2026-11-19 9:59 ` [PATCH pve-manager v5 01/10] fix #6144: api: nodes: add endpoint to set " Dominik Rusovac
2026-11-19 9:59 ` [PATCH pve-manager v5 02/10] fix #6144: ui: ha: status: allow toggling " Dominik Rusovac
@ 2026-11-19 9:59 ` Dominik Rusovac
2026-11-19 9:59 ` [PATCH pve-ha-manager v5 04/10] test: add test for ignoring maintenance mode request Dominik Rusovac
` (7 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: Dominik Rusovac @ 2026-11-19 9:59 UTC (permalink / raw)
To: pve-devel
the CRM only processes a node maintenance request on its next round, so
until then, there is no feedback whether the request is still being
processed, and the menu still offers the same action again
show a spinner on the lrm entry of the node and keep the menu disabled
for it while the request is queued, like it is done when disarming HA.
this also ends the spinner if the CRM ignores the request, e.g. if the
node went offline in the meantime
as the request might not be reported as queued by a status load that was
already running, the node is also considered pending from the successful
request until a status load that started afterwards
Signed-off-by: Dominik Rusovac <d.rusovac@proxmox.com>
---
Notes:
new in v5, changes w.r.t. v4:
* show the spinner while the request is queued for the CRM, so that
it also ends if the CRM ignores the request, as pointed out in the
v4 review
* only initialize the per-node pending state in initComponent
www/manager6/ha/Status.js | 11 +++++++++-
www/manager6/ha/StatusView.js | 41 ++++++++++++++++++++++++++++++++++-
2 files changed, 50 insertions(+), 2 deletions(-)
diff --git a/www/manager6/ha/Status.js b/www/manager6/ha/Status.js
index 42e9aefb..44f95d61 100644
--- a/www/manager6/ha/Status.js
+++ b/www/manager6/ha/Status.js
@@ -14,10 +14,13 @@ Ext.define('PVE.ha.Status', {
selectedLrmNode: null,
// maintenance state of the selected LRM node
maintenanceRequested: false,
+ maintenancePending: false,
},
formulas: {
noMaintenanceAction: (get) =>
- !get('selectedLrmNode') || !Ext.state.Manager.get('GuiCap').nodes['Sys.Modify'],
+ !get('selectedLrmNode') ||
+ get('maintenancePending') ||
+ !Ext.state.Manager.get('GuiCap').nodes['Sys.Modify'],
// the CRM ignores requests to enable maintenance mode while HA is disarmed
noMaintenanceEnable: (get) => get('maintenanceRequested') || get('haDisarmed'),
},
@@ -70,6 +73,7 @@ Ext.define('PVE.ha.Status', {
vm.set('selectedLrmNode', isLrm ? rec.get('node') : null);
vm.set('maintenanceRequested', isLrm && rec.get('maintenance-requested'));
+ vm.set('maintenancePending', isLrm && sv.isMaintenancePending(rec));
},
handleMaintenanceButton: function (menuItem) {
@@ -115,6 +119,11 @@ Ext.define('PVE.ha.Status', {
url: `/nodes/${node}/maintenance`,
params: { 'maintenance-mode': menuItem.maintenanceMode },
method: 'PUT',
+ success: function () {
+ let sv = me.getView().down('pveHAStatusView');
+ sv.setMaintenancePending(node);
+ me.updateMaintenanceState();
+ },
failure: (response) => Ext.Msg.alert(gettext('Error'), response.htmlStatus),
});
});
diff --git a/www/manager6/ha/StatusView.js b/www/manager6/ha/StatusView.js
index d08534b4..cf8677e4 100644
--- a/www/manager6/ha/StatusView.js
+++ b/www/manager6/ha/StatusView.js
@@ -23,6 +23,19 @@ Ext.define(
me.getView().refresh();
},
+ setMaintenancePending: function (node) {
+ let me = this;
+ me.maintenanceRequestTime[node] = Date.now();
+ me.getView().refresh();
+ },
+
+ isMaintenancePending: function (rec) {
+ return (
+ rec.get('maintenance-request-queued') ||
+ this.maintenanceRequestTime[rec.get('node')] !== undefined
+ );
+ },
+
initComponent: function () {
var me = this;
@@ -30,6 +43,10 @@ Ext.define(
throw 'no rstore given';
}
+ // maps node names to the time of their last maintenance request, until a status load
+ // that started afterwards reports whether the request is still queued for the CRM
+ me.maintenanceRequestTime = {};
+
Proxmox.Utils.monStoreErrors(me, me.rstore);
var store = Ext.create('Proxmox.data.DiffStore', {
@@ -51,7 +68,22 @@ Ext.define(
},
});
+ me.rstore.on('beforeload', () => {
+ me.lastLoadStart = Date.now();
+ });
+
me.rstore.on('load', function () {
+ let pendingCleared = false;
+ for (const [node, requestTime] of Object.entries(me.maintenanceRequestTime)) {
+ if (requestTime < me.lastLoadStart) {
+ delete me.maintenanceRequestTime[node];
+ pendingCleared = true;
+ }
+ }
+ if (pendingCleared) {
+ me.getView().refresh();
+ }
+
let fencing = store.findRecord('type', 'fencing');
let disarmed = fencing && fencing.get('armed-state') === 'disarmed';
@@ -76,7 +108,10 @@ Ext.define(
flex: 1,
dataIndex: 'status',
renderer: function (value, _metaData, rec) {
- if (rec.data.type === 'fencing' && me.pending) {
+ if (
+ (rec.data.type === 'fencing' && me.pending) ||
+ (rec.data.type === 'lrm' && me.isMaintenancePending(rec))
+ ) {
return value + ' <i class="fa fa-spinner fa-spin"></i>';
}
return value;
@@ -136,6 +171,10 @@ Ext.define(
name: 'maintenance-requested',
type: 'boolean',
},
+ {
+ name: 'maintenance-request-queued',
+ type: 'boolean',
+ },
],
idProperty: 'id',
});
--
2.47.3
^ permalink raw reply related [flat|nested] 12+ messages in thread* [PATCH pve-ha-manager v5 04/10] test: add test for ignoring maintenance mode request
2026-11-19 9:59 [PATCH-SERIES docs/ha-manager/manager v5 00/10] fix #6144: add api and ui for node maintenance mode Dominik Rusovac
` (2 preceding siblings ...)
2026-11-19 9:59 ` [PATCH pve-manager v5 03/10] ui: ha: status: show spinner while maintenance mode request is queued Dominik Rusovac
@ 2026-11-19 9:59 ` Dominik Rusovac
2026-11-19 9:59 ` [PATCH pve-ha-manager v5 05/10] manager: standardize log messages for " Dominik Rusovac
` (6 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: Dominik Rusovac @ 2026-11-19 9:59 UTC (permalink / raw)
To: pve-devel
none of the existing tests covers the cases where the CRM ignores a
request to enable maintenance mode for a node that is already in
maintenance mode or while the HA stack is disarmed
Signed-off-by: Dominik Rusovac <d.rusovac@proxmox.com>
---
Notes:
new in v5
src/test/test-crm-maintenance-request6/README | 3 +
.../test-crm-maintenance-request6/cmdlist | 9 ++
.../hardware_status | 5 ++
.../test-crm-maintenance-request6/log.expect | 82 +++++++++++++++++++
.../manager_status | 1 +
.../service_config | 5 ++
6 files changed, 105 insertions(+)
create mode 100644 src/test/test-crm-maintenance-request6/README
create mode 100644 src/test/test-crm-maintenance-request6/cmdlist
create mode 100644 src/test/test-crm-maintenance-request6/hardware_status
create mode 100644 src/test/test-crm-maintenance-request6/log.expect
create mode 100644 src/test/test-crm-maintenance-request6/manager_status
create mode 100644 src/test/test-crm-maintenance-request6/service_config
diff --git a/src/test/test-crm-maintenance-request6/README b/src/test/test-crm-maintenance-request6/README
new file mode 100644
index 0000000..a7a510f
--- /dev/null
+++ b/src/test/test-crm-maintenance-request6/README
@@ -0,0 +1,3 @@
+Test that the CRM ignores requests to enable the maintenance mode for a node
+that is already in maintenance mode, and while the HA stack is disarmed,
+respectively.
diff --git a/src/test/test-crm-maintenance-request6/cmdlist b/src/test/test-crm-maintenance-request6/cmdlist
new file mode 100644
index 0000000..574ee6a
--- /dev/null
+++ b/src/test/test-crm-maintenance-request6/cmdlist
@@ -0,0 +1,9 @@
+[
+ [ "power node1 on", "power node2 on", "power node3 on"],
+ [ "crm node3 enable-node-maintenance" ],
+ [ "crm node3 enable-node-maintenance" ],
+ [ "crm node1 disarm-ha freeze" ],
+ [ "crm node2 enable-node-maintenance" ],
+ [ "crm node1 arm-ha" ],
+ [ "crm node3 disable-node-maintenance" ]
+]
diff --git a/src/test/test-crm-maintenance-request6/hardware_status b/src/test/test-crm-maintenance-request6/hardware_status
new file mode 100644
index 0000000..451beb1
--- /dev/null
+++ b/src/test/test-crm-maintenance-request6/hardware_status
@@ -0,0 +1,5 @@
+{
+ "node1": { "power": "off", "network": "off" },
+ "node2": { "power": "off", "network": "off" },
+ "node3": { "power": "off", "network": "off" }
+}
diff --git a/src/test/test-crm-maintenance-request6/log.expect b/src/test/test-crm-maintenance-request6/log.expect
new file mode 100644
index 0000000..1db9138
--- /dev/null
+++ b/src/test/test-crm-maintenance-request6/log.expect
@@ -0,0 +1,82 @@
+info 0 hardware: starting simulation
+info 20 cmdlist: execute power node1 on
+info 20 node1/crm: status change startup => wait_for_quorum
+info 20 node1/lrm: status change startup => wait_for_agent_lock
+info 20 cmdlist: execute power node2 on
+info 20 node2/crm: status change startup => wait_for_quorum
+info 20 node2/lrm: status change startup => wait_for_agent_lock
+info 20 cmdlist: execute power node3 on
+info 20 node3/crm: status change startup => wait_for_quorum
+info 20 node3/lrm: status change startup => wait_for_agent_lock
+info 20 node1/crm: got lock 'ha_manager_lock'
+info 20 node1/crm: status change wait_for_quorum => master
+info 20 node1/crm: node 'node1': state changed from 'unknown' => 'online'
+info 20 node1/crm: node 'node2': state changed from 'unknown' => 'online'
+info 20 node1/crm: node 'node3': state changed from 'unknown' => 'online'
+info 20 node1/crm: adding new service 'vm:101' on node 'node1'
+info 20 node1/crm: adding new service 'vm:102' on node 'node2'
+info 20 node1/crm: adding new service 'vm:103' on node 'node3'
+info 20 node1/crm: service 'vm:101': state changed from 'request_start' to 'started' (node = node1)
+info 20 node1/crm: service 'vm:103': state changed from 'request_start' to 'started' (node = node3)
+info 21 node1/lrm: got lock 'ha_agent_node1_lock'
+info 21 node1/lrm: status change wait_for_agent_lock => active
+info 21 node1/lrm: starting service vm:101
+info 21 node1/lrm: service status vm:101 started
+info 22 node2/crm: status change wait_for_quorum => slave
+info 23 node2/lrm: got lock 'ha_agent_node2_lock'
+info 23 node2/lrm: status change wait_for_agent_lock => active
+info 24 node3/crm: status change wait_for_quorum => slave
+info 25 node3/lrm: got lock 'ha_agent_node3_lock'
+info 25 node3/lrm: status change wait_for_agent_lock => active
+info 25 node3/lrm: starting service vm:103
+info 25 node3/lrm: service status vm:103 started
+info 40 node1/crm: service 'vm:102': state changed from 'request_stop' to 'stopped'
+info 120 cmdlist: execute crm node3 enable-node-maintenance
+info 125 node3/lrm: status change active => maintenance
+info 140 node1/crm: node 'node3': state changed from 'online' => 'maintenance'
+info 140 node1/crm: migrate service 'vm:103' to node 'node2' (running)
+info 140 node1/crm: service 'vm:103': state changed from 'started' to 'migrate' (node = node3, target = node2)
+info 145 node3/lrm: service vm:103 - start migrate to node 'node2'
+info 145 node3/lrm: service vm:103 - end migrate to node 'node2'
+info 160 node1/crm: service 'vm:103': state changed from 'migrate' to 'started' (node = node2)
+info 163 node2/lrm: starting service vm:103
+info 163 node2/lrm: service status vm:103 started
+info 220 cmdlist: execute crm node3 enable-node-maintenance
+info 220 node1/crm: ignoring crm command - node node3 is already in maintenance state
+info 320 cmdlist: execute crm node1 disarm-ha freeze
+info 320 node1/crm: got crm command: disarm-ha freeze
+info 320 node1/crm: disarm: freezing service 'vm:101' (was 'started')
+info 320 node1/crm: disarm: freezing service 'vm:102' (was 'stopped')
+info 320 node1/crm: disarm: freezing service 'vm:103' (was 'started')
+info 321 node1/lrm: HA disarm requested, releasing agent lock and watchdog
+info 321 node1/lrm: status change active => wait_for_agent_lock
+info 323 node2/lrm: HA disarm requested, releasing agent lock and watchdog
+info 323 node2/lrm: status change active => wait_for_agent_lock
+info 340 node1/crm: all LRMs disarmed, HA stack is now fully disarmed
+info 340 node1/crm: HA stack fully disarmed, releasing CRM watchdog
+info 420 cmdlist: execute crm node2 enable-node-maintenance
+warn 420 node1/crm: ignoring maintenance command for node node2 - HA stack is disarmed
+info 520 cmdlist: execute crm node1 arm-ha
+info 520 node1/crm: got crm command: arm-ha
+info 520 node1/crm: re-arming HA stack
+info 540 node1/crm: service 'vm:101': state changed from 'freeze' to 'started'
+info 540 node1/crm: service 'vm:102': state changed from 'freeze' to 'request_stop'
+info 540 node1/crm: service 'vm:103': state changed from 'freeze' to 'started'
+info 541 node1/lrm: got lock 'ha_agent_node1_lock'
+info 541 node1/lrm: status change wait_for_agent_lock => active
+info 543 node2/lrm: got lock 'ha_agent_node2_lock'
+info 543 node2/lrm: status change wait_for_agent_lock => active
+info 560 node1/crm: service 'vm:102': state changed from 'request_stop' to 'stopped'
+info 620 cmdlist: execute crm node3 disable-node-maintenance
+info 625 node3/lrm: got lock 'ha_agent_node3_lock'
+info 625 node3/lrm: status change maintenance => active
+info 640 node1/crm: node 'node3': state changed from 'maintenance' => 'online'
+info 640 node1/crm: moving service 'vm:103' back to 'node3', node came back from maintenance.
+info 640 node1/crm: migrate service 'vm:103' to node 'node3' (running)
+info 640 node1/crm: service 'vm:103': state changed from 'started' to 'migrate' (node = node2, target = node3)
+info 643 node2/lrm: service vm:103 - start migrate to node 'node3'
+info 643 node2/lrm: service vm:103 - end migrate to node 'node3'
+info 660 node1/crm: service 'vm:103': state changed from 'migrate' to 'started' (node = node3)
+info 665 node3/lrm: starting service vm:103
+info 665 node3/lrm: service status vm:103 started
+info 1220 hardware: exit simulation - done
diff --git a/src/test/test-crm-maintenance-request6/manager_status b/src/test/test-crm-maintenance-request6/manager_status
new file mode 100644
index 0000000..0967ef4
--- /dev/null
+++ b/src/test/test-crm-maintenance-request6/manager_status
@@ -0,0 +1 @@
+{}
diff --git a/src/test/test-crm-maintenance-request6/service_config b/src/test/test-crm-maintenance-request6/service_config
new file mode 100644
index 0000000..c943ba6
--- /dev/null
+++ b/src/test/test-crm-maintenance-request6/service_config
@@ -0,0 +1,5 @@
+{
+ "vm:101": { "node": "node1", "state": "started" },
+ "vm:102": { "node": "node2" },
+ "vm:103": { "node": "node3", "state": "started" }
+}
\ No newline at end of file
--
2.47.3
^ permalink raw reply related [flat|nested] 12+ messages in thread* [PATCH pve-ha-manager v5 05/10] manager: standardize log messages for maintenance mode request
2026-11-19 9:59 [PATCH-SERIES docs/ha-manager/manager v5 00/10] fix #6144: add api and ui for node maintenance mode Dominik Rusovac
` (3 preceding siblings ...)
2026-11-19 9:59 ` [PATCH pve-ha-manager v5 04/10] test: add test for ignoring maintenance mode request Dominik Rusovac
@ 2026-11-19 9:59 ` Dominik Rusovac
2026-11-19 9:59 ` [PATCH pve-ha-manager v5 06/10] manager: factor out check " Dominik Rusovac
` (5 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: Dominik Rusovac @ 2026-11-19 9:59 UTC (permalink / raw)
To: pve-devel
log messages for maintenance mode request now share a format describing
what the command does, and no longer depend on cmd variable
adapt the affected test logs accordingly
Signed-off-by: Dominik Rusovac <d.rusovac@proxmox.com>
---
Notes:
new in v5
src/PVE/HA/Manager.pm | 6 +++---
src/test/test-crm-maintenance-request2/log.expect | 2 +-
src/test/test-crm-maintenance-request6/log.expect | 4 ++--
3 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/src/PVE/HA/Manager.pm b/src/PVE/HA/Manager.pm
index 5c21fbe..90ef89d 100644
--- a/src/PVE/HA/Manager.pm
+++ b/src/PVE/HA/Manager.pm
@@ -794,17 +794,17 @@ sub update_crm_commands {
if ($ms->{disarm}) {
$haenv->log(
'warning',
- "ignoring maintenance command for node $node - HA stack is disarmed",
+ "ignoring maintenance mode request for node '$node' - HA stack is disarmed",
);
} elsif ($state eq 'online') {
$ms->{node_request}->{$node}->{maintenance} = 1;
} elsif ($state eq 'maintenance') {
$haenv->log(
'info',
- "ignoring crm command - node $node is already in maintenance state",
+ "ignoring maintenance mode request for node '$node' - node is already in maintenance state",
);
} else {
- $haenv->log('err', "crm command error - node not online: $cmd");
+ $haenv->log('err', "maintenance mode request error - node '$node' not online");
}
} elsif ($cmd =~ m/^disable-node-maintenance\s+(\S+)$/) {
my $node = $1;
diff --git a/src/test/test-crm-maintenance-request2/log.expect b/src/test/test-crm-maintenance-request2/log.expect
index c8d2563..abf86f8 100644
--- a/src/test/test-crm-maintenance-request2/log.expect
+++ b/src/test/test-crm-maintenance-request2/log.expect
@@ -43,7 +43,7 @@ info 165 node3/crm: killed by poweroff
info 166 node3/lrm: killed by poweroff
info 166 hardware: server 'node3' stopped by poweroff (watchdog)
info 220 cmdlist: execute crm node3 enable-node-maintenance
-err 220 node1/crm: crm command error - node not online: enable-node-maintenance node3
+err 220 node1/crm: maintenance mode request error - node 'node3' not online
info 240 node1/crm: got lock 'ha_agent_node3_lock'
info 240 node1/crm: fencing: acknowledged - got agent lock for node 'node3'
info 240 node1/crm: node 'node3': state changed from 'fence' => 'unknown'
diff --git a/src/test/test-crm-maintenance-request6/log.expect b/src/test/test-crm-maintenance-request6/log.expect
index 1db9138..7ca0d61 100644
--- a/src/test/test-crm-maintenance-request6/log.expect
+++ b/src/test/test-crm-maintenance-request6/log.expect
@@ -42,7 +42,7 @@ info 160 node1/crm: service 'vm:103': state changed from 'migrate' to 'sta
info 163 node2/lrm: starting service vm:103
info 163 node2/lrm: service status vm:103 started
info 220 cmdlist: execute crm node3 enable-node-maintenance
-info 220 node1/crm: ignoring crm command - node node3 is already in maintenance state
+info 220 node1/crm: ignoring maintenance mode request for node 'node3' - node is already in maintenance state
info 320 cmdlist: execute crm node1 disarm-ha freeze
info 320 node1/crm: got crm command: disarm-ha freeze
info 320 node1/crm: disarm: freezing service 'vm:101' (was 'started')
@@ -55,7 +55,7 @@ info 323 node2/lrm: status change active => wait_for_agent_lock
info 340 node1/crm: all LRMs disarmed, HA stack is now fully disarmed
info 340 node1/crm: HA stack fully disarmed, releasing CRM watchdog
info 420 cmdlist: execute crm node2 enable-node-maintenance
-warn 420 node1/crm: ignoring maintenance command for node node2 - HA stack is disarmed
+warn 420 node1/crm: ignoring maintenance mode request for node 'node2' - HA stack is disarmed
info 520 cmdlist: execute crm node1 arm-ha
info 520 node1/crm: got crm command: arm-ha
info 520 node1/crm: re-arming HA stack
--
2.47.3
^ permalink raw reply related [flat|nested] 12+ messages in thread* [PATCH pve-ha-manager v5 06/10] manager: factor out check for maintenance mode request
2026-11-19 9:59 [PATCH-SERIES docs/ha-manager/manager v5 00/10] fix #6144: add api and ui for node maintenance mode Dominik Rusovac
` (4 preceding siblings ...)
2026-11-19 9:59 ` [PATCH pve-ha-manager v5 05/10] manager: standardize log messages for " Dominik Rusovac
@ 2026-11-19 9:59 ` Dominik Rusovac
2026-11-19 9:59 ` [PATCH pve-ha-manager v5 07/10] cli: reject maintenance mode requests rejected by crm Dominik Rusovac
` (4 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: Dominik Rusovac @ 2026-11-19 9:59 UTC (permalink / raw)
To: pve-devel
move the reasons why the crm ignores a maintenance mode request into a
helper, so that the CLI can use the same logic to reject such requests
in advance, instead of queuing a crm command that is only ignored with a
log message
Signed-off-by: Dominik Rusovac <d.rusovac@proxmox.com>
---
Notes:
new in v5:
* uses signature for function in Helpers to be consistent with rest of
file
src/PVE/HA/Config.pm | 9 +++++++++
src/PVE/HA/Helpers.pm | 21 +++++++++++++++++++++
src/PVE/HA/Manager.pm | 19 ++++++-------------
3 files changed, 36 insertions(+), 13 deletions(-)
diff --git a/src/PVE/HA/Config.pm b/src/PVE/HA/Config.pm
index bee1ac7..b3ab77b 100644
--- a/src/PVE/HA/Config.pm
+++ b/src/PVE/HA/Config.pm
@@ -423,6 +423,15 @@ sub get_resource_motion_info {
return ($dependent_resources, $blocking_resources_by_node);
}
+sub check_enable_maintenance {
+ my ($node, $log_fn) = @_;
+
+ my $ms = read_manager_status();
+ my $ns = ($ms->{node_status} // {})->{$node} // 'unknown';
+
+ return PVE::HA::Helpers::check_enable_maintenance($log_fn, $node, $ns, $ms);
+}
+
# graceful, as long as locking + cfs_write works
sub delete_service_from_config {
my ($sid, $purge) = @_;
diff --git a/src/PVE/HA/Helpers.pm b/src/PVE/HA/Helpers.pm
index a58b1e1..ba69093 100644
--- a/src/PVE/HA/Helpers.pm
+++ b/src/PVE/HA/Helpers.pm
@@ -60,4 +60,25 @@ sub get_resource_motion_info($ss, $sid, $cd, $online_nodes, $compiled_rules) {
return ($dependent_resources, $blocking_resources_by_node);
}
+sub check_enable_maintenance($log_fn, $node, $node_state, $ms) {
+ if ($ms->{disarm}) {
+ $log_fn->(
+ 'warning',
+ "ignoring maintenance mode request for node '$node' - HA stack is disarmed",
+ );
+ return 0;
+ } elsif ($node_state eq 'maintenance') {
+ $log_fn->(
+ 'info',
+ "ignoring maintenance mode request for node '$node' - node is already in maintenance state",
+ );
+ return 0;
+ } elsif ($node_state ne 'online') {
+ $log_fn->('err', "maintenance mode request error - node '$node' not online");
+ return 0;
+ }
+
+ return 1;
+}
+
1;
diff --git a/src/PVE/HA/Manager.pm b/src/PVE/HA/Manager.pm
index 90ef89d..c012750 100644
--- a/src/PVE/HA/Manager.pm
+++ b/src/PVE/HA/Manager.pm
@@ -770,6 +770,11 @@ sub update_crm_commands {
my $cmdlist = $haenv->read_crm_commands();
+ my $log_fn = sub {
+ my ($level, $msg) = @_;
+ $haenv->log($level, $msg);
+ };
+
foreach my $cmd (split(/\n/, $cmdlist)) {
chomp $cmd;
@@ -791,20 +796,8 @@ sub update_crm_commands {
my $node = $1;
my $state = $ns->get_node_state($node);
- if ($ms->{disarm}) {
- $haenv->log(
- 'warning',
- "ignoring maintenance mode request for node '$node' - HA stack is disarmed",
- );
- } elsif ($state eq 'online') {
+ if (PVE::HA::Helpers::check_enable_maintenance($log_fn, $node, $state, $ms)) {
$ms->{node_request}->{$node}->{maintenance} = 1;
- } elsif ($state eq 'maintenance') {
- $haenv->log(
- 'info',
- "ignoring maintenance mode request for node '$node' - node is already in maintenance state",
- );
- } else {
- $haenv->log('err', "maintenance mode request error - node '$node' not online");
}
} elsif ($cmd =~ m/^disable-node-maintenance\s+(\S+)$/) {
my $node = $1;
--
2.47.3
^ permalink raw reply related [flat|nested] 12+ messages in thread* [PATCH pve-ha-manager v5 07/10] cli: reject maintenance mode requests rejected by crm
2026-11-19 9:59 [PATCH-SERIES docs/ha-manager/manager v5 00/10] fix #6144: add api and ui for node maintenance mode Dominik Rusovac
` (5 preceding siblings ...)
2026-11-19 9:59 ` [PATCH pve-ha-manager v5 06/10] manager: factor out check " Dominik Rusovac
@ 2026-11-19 9:59 ` Dominik Rusovac
2026-11-19 9:59 ` [PATCH pve-ha-manager v5 08/10] api: status: report maintenance mode request for lrm entries Dominik Rusovac
` (3 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: Dominik Rusovac @ 2026-11-19 9:59 UTC (permalink / raw)
To: pve-devel
the CLI queued maintenance mode requests with no checks, so users only
learned from the CRM's log that a request was rejected
reject those before queuing the CRM command using the same logic as the
CRM. this only concerns requests to enable maintenance mode, as the CRM
always processes requests to disable it
requests for nodes unknown to HA are rejected as well, since maintenance
mode only affects HA managed resources and the queued command would make
an idle CRM become active
dying in the log function is supposed to forward the reason for
rejection to callers
deref node parameter once
Signed-off-by: Dominik Rusovac <d.rusovac@proxmox.com>
---
Notes:
new in v5
src/PVE/CLI/ha_manager.pm | 22 +++++++++++++++++++---
1 file changed, 19 insertions(+), 3 deletions(-)
diff --git a/src/PVE/CLI/ha_manager.pm b/src/PVE/CLI/ha_manager.pm
index 6625de6..74f18d9 100644
--- a/src/PVE/CLI/ha_manager.pm
+++ b/src/PVE/CLI/ha_manager.pm
@@ -138,10 +138,26 @@ __PACKAGE__->register_method({
code => sub {
my ($param) = @_;
- PVE::Cluster::check_node_exists($param->{node});
+ my $node = $param->{node};
+ PVE::Cluster::check_node_exists($node);
+
+ # maintenance mode only affects HA managed resources and queuing the CRM command for a
+ # node unknown to HA would make an idle CRM become active for nothing
+ my $node_status = PVE::HA::Config::read_manager_status()->{node_status} // {};
+ die "node '$node' is not part of HA\n" if !defined($node_status->{$node});
+
+ my $disable = $param->{disable};
+ if (!$disable) {
+ my $log_fn = sub {
+ my ($level, $msg) = @_;
+ die "$msg\n";
+ };
+
+ PVE::HA::Config::check_enable_maintenance($node, $log_fn);
+ }
- my $cmd = $param->{disable} ? 'disable-node-maintenance' : 'enable-node-maintenance';
- PVE::HA::Config::queue_crm_commands("$cmd $param->{node}");
+ my $cmd = $disable ? 'disable-node-maintenance' : 'enable-node-maintenance';
+ PVE::HA::Config::queue_crm_commands("$cmd $node");
return undef;
},
--
2.47.3
^ permalink raw reply related [flat|nested] 12+ messages in thread* [PATCH pve-ha-manager v5 08/10] api: status: report maintenance mode request for lrm entries
2026-11-19 9:59 [PATCH-SERIES docs/ha-manager/manager v5 00/10] fix #6144: add api and ui for node maintenance mode Dominik Rusovac
` (6 preceding siblings ...)
2026-11-19 9:59 ` [PATCH pve-ha-manager v5 07/10] cli: reject maintenance mode requests rejected by crm Dominik Rusovac
@ 2026-11-19 9:59 ` Dominik Rusovac
2026-11-19 9:59 ` [PATCH pve-ha-manager v5 09/10] api: status: report queued " Dominik Rusovac
` (2 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: Dominik Rusovac @ 2026-11-19 9:59 UTC (permalink / raw)
To: pve-devel; +Cc: Thomas Skinner
expose whether maintenance mode was manually requested for a node, as
recorded by the CRM
Originally-by: Thomas Skinner <thomas@atskinner.net>
Signed-off-by: Dominik Rusovac <d.rusovac@proxmox.com>
Reviewed-by: Daniel Kral <d.kral@proxmox.com>
---
Notes:
changes since v4:
* reword commit subject
src/PVE/API2/HA/Status.pm | 10 ++++++++++
1 file changed, 10 insertions(+)
diff --git a/src/PVE/API2/HA/Status.pm b/src/PVE/API2/HA/Status.pm
index aecef0f..2622ec1 100644
--- a/src/PVE/API2/HA/Status.pm
+++ b/src/PVE/API2/HA/Status.pm
@@ -107,6 +107,12 @@ __PACKAGE__->register_method({
type => "integer",
optional => 1,
},
+ 'maintenance-requested' => {
+ description => "For type 'lrm'. Whether maintenance mode was manually"
+ . " requested for the node.",
+ type => "boolean",
+ optional => 1,
+ },
crm_state => {
description => "For type 'service'. Service state as seen by the CRM.",
type => "string",
@@ -265,6 +271,8 @@ __PACKAGE__->register_method({
PVE::HA::Tools::count_active_services($status->{service_status}, $node);
my $lrm_status = PVE::HA::Config::read_lrm_status($node);
my $id = "lrm:$node";
+ my $node_request = $status->{node_request} // {};
+ my $maintenance_requested = $node_request->{$node}->{maintenance} ? 1 : 0;
if (!$lrm_status->{timestamp}) {
push @$res,
{
@@ -272,6 +280,7 @@ __PACKAGE__->register_method({
type => 'lrm',
node => $node,
status => "$node (unable to read lrm status)",
+ 'maintenance-requested' => $maintenance_requested,
};
} else {
my $status_str = &$timestamp_to_status($ctime, $lrm_status->{timestamp});
@@ -315,6 +324,7 @@ __PACKAGE__->register_method({
node => $node,
status => $status_text,
timestamp => $lrm_status->{timestamp},
+ 'maintenance-requested' => $maintenance_requested,
};
}
}
--
2.47.3
^ permalink raw reply related [flat|nested] 12+ messages in thread* [PATCH pve-ha-manager v5 09/10] api: status: report queued maintenance mode request for lrm entries
2026-11-19 9:59 [PATCH-SERIES docs/ha-manager/manager v5 00/10] fix #6144: add api and ui for node maintenance mode Dominik Rusovac
` (7 preceding siblings ...)
2026-11-19 9:59 ` [PATCH pve-ha-manager v5 08/10] api: status: report maintenance mode request for lrm entries Dominik Rusovac
@ 2026-11-19 9:59 ` Dominik Rusovac
2026-11-19 9:59 ` [PATCH pve-docs v5 10/10] ha-manager: maintenance mode: mention web ui menu Dominik Rusovac
2026-11-19 22:28 ` applied: [PATCH-SERIES docs/ha-manager/manager v5 00/10] fix #6144: add api and ui for node maintenance mode Thomas Lamprecht
10 siblings, 0 replies; 12+ messages in thread
From: Dominik Rusovac @ 2026-11-19 9:59 UTC (permalink / raw)
To: pve-devel
expose whether a request to enable/disable maintenance mode for a node
is still queued, so that clients can tell when the CRM processed it; no
matter if CRM accepted or ignored the request
peeking into the queued CRM commands reads them without consuming them
Signed-off-by: Dominik Rusovac <d.rusovac@proxmox.com>
---
Notes:
new in v5:
* towards fixing the bug mentioned in the review of v4
* only needed for the spinner, can be left out together with it
src/PVE/API2/HA/Status.pm | 16 ++++++++++++++++
src/PVE/HA/Config.pm | 4 ++++
2 files changed, 20 insertions(+)
diff --git a/src/PVE/API2/HA/Status.pm b/src/PVE/API2/HA/Status.pm
index 2622ec1..e3a4e16 100644
--- a/src/PVE/API2/HA/Status.pm
+++ b/src/PVE/API2/HA/Status.pm
@@ -113,6 +113,13 @@ __PACKAGE__->register_method({
type => "boolean",
optional => 1,
},
+ 'maintenance-request-queued' => {
+ description => "For type 'lrm'. Whether a request to enable or disable"
+ . " maintenance mode for the node is queued and not yet processed by"
+ . " the CRM.",
+ type => "boolean",
+ optional => 1,
+ },
crm_state => {
description => "For type 'service'. Service state as seen by the CRM.",
type => "string",
@@ -266,6 +273,12 @@ __PACKAGE__->register_method({
};
}
+ my $maintenance_queued_nodes = {};
+ for my $cmd (split(/\n/, PVE::HA::Config::peek_crm_commands())) {
+ next if $cmd !~ m/^(?:enable|disable)-node-maintenance\s+(\S+)$/;
+ $maintenance_queued_nodes->{$1} = 1;
+ }
+
foreach my $node (sort keys %{ $status->{node_status} }) {
my $active_count =
PVE::HA::Tools::count_active_services($status->{service_status}, $node);
@@ -273,6 +286,7 @@ __PACKAGE__->register_method({
my $id = "lrm:$node";
my $node_request = $status->{node_request} // {};
my $maintenance_requested = $node_request->{$node}->{maintenance} ? 1 : 0;
+ my $maintenance_queued = $maintenance_queued_nodes->{$node} ? 1 : 0;
if (!$lrm_status->{timestamp}) {
push @$res,
{
@@ -281,6 +295,7 @@ __PACKAGE__->register_method({
node => $node,
status => "$node (unable to read lrm status)",
'maintenance-requested' => $maintenance_requested,
+ 'maintenance-request-queued' => $maintenance_queued,
};
} else {
my $status_str = &$timestamp_to_status($ctime, $lrm_status->{timestamp});
@@ -325,6 +340,7 @@ __PACKAGE__->register_method({
status => $status_text,
timestamp => $lrm_status->{timestamp},
'maintenance-requested' => $maintenance_requested,
+ 'maintenance-request-queued' => $maintenance_queued,
};
}
}
diff --git a/src/PVE/HA/Config.pm b/src/PVE/HA/Config.pm
index b3ab77b..4f02d5f 100644
--- a/src/PVE/HA/Config.pm
+++ b/src/PVE/HA/Config.pm
@@ -341,6 +341,10 @@ sub any_pending_crm_command {
return defined($data) && length $data;
}
+sub peek_crm_commands {
+ return cfs_read_file($crm_commands_filename) // '';
+}
+
sub read_crm_commands {
my $code = sub {
--
2.47.3
^ permalink raw reply related [flat|nested] 12+ messages in thread* [PATCH pve-docs v5 10/10] ha-manager: maintenance mode: mention web ui menu
2026-11-19 9:59 [PATCH-SERIES docs/ha-manager/manager v5 00/10] fix #6144: add api and ui for node maintenance mode Dominik Rusovac
` (8 preceding siblings ...)
2026-11-19 9:59 ` [PATCH pve-ha-manager v5 09/10] api: status: report queued " Dominik Rusovac
@ 2026-11-19 9:59 ` Dominik Rusovac
2026-11-19 22:28 ` applied: [PATCH-SERIES docs/ha-manager/manager v5 00/10] fix #6144: add api and ui for node maintenance mode Thomas Lamprecht
10 siblings, 0 replies; 12+ messages in thread
From: Dominik Rusovac @ 2026-11-19 9:59 UTC (permalink / raw)
To: pve-devel; +Cc: Thomas Skinner
maintenance mode can now also be enabled/disabled from the HA status
panel
describe where to find it and that it cannot be enabled while HA is
disarmed
fix the "enabled or disable" typo in the rewritten sentence
Originally-by: Thomas Skinner <thomas@atskinner.net>
Signed-off-by: Dominik Rusovac <d.rusovac@proxmox.com>
---
Notes:
no changes since v4
ha-manager.adoc | 15 +++++++++------
1 file changed, 9 insertions(+), 6 deletions(-)
diff --git a/ha-manager.adoc b/ha-manager.adoc
index 95ab799..4962bb9 100644
--- a/ha-manager.adoc
+++ b/ha-manager.adoc
@@ -1199,8 +1199,10 @@ During each migration, the original node will be recorded in the HA managers'
state, so that the service can be moved back again automatically once the
maintenance mode is disabled and the node is back online.
-Currently you can enabled or disable the maintenance mode using the ha-manager
-CLI tool.
+You can enable or disable the maintenance mode using the web interface or the
+`ha-manager` CLI tool. In the web interface, select the `lrm` entry of the node
+in the status panel under __Datacenter -> HA__ and use the *Node Maintenance*
+menu. Maintenance mode cannot be enabled while HA is disarmed.
.Enabling maintenance mode for a node
----
@@ -1225,8 +1227,8 @@ out for a log line like: `pve-ha-lrm[PID]: watchdog closed (disabled)` to know
when the node finished its transition into the maintenance mode.
NOTE: The manual maintenance mode is not automatically deleted on node reboot,
-but only if it is either manually deactivated using the `ha-manager` CLI or if
-the manager-status is manually cleared.
+but only if it is either manually deactivated or if the manager-status is
+manually cleared.
.Disabling maintenance mode for a node
----
@@ -1234,8 +1236,9 @@ the manager-status is manually cleared.
----
The process of disabling the manual maintenance mode is similar to enabling it.
-Using the `ha-manager` CLI command shown above will queue a CRM command that,
-once processed, marks the respective LRM node as available again.
+Using the web interface or the `ha-manager` CLI command shown above will queue a
+CRM command that, once processed, marks the respective LRM node as available
+again.
If you deactivate the maintenance mode, all services that were on the node when
the maintenance mode was activated will be moved back.
--
2.47.3
^ permalink raw reply related [flat|nested] 12+ messages in thread* applied: [PATCH-SERIES docs/ha-manager/manager v5 00/10] fix #6144: add api and ui for node maintenance mode
2026-11-19 9:59 [PATCH-SERIES docs/ha-manager/manager v5 00/10] fix #6144: add api and ui for node maintenance mode Dominik Rusovac
` (9 preceding siblings ...)
2026-11-19 9:59 ` [PATCH pve-docs v5 10/10] ha-manager: maintenance mode: mention web ui menu Dominik Rusovac
@ 2026-11-19 22:28 ` Thomas Lamprecht
10 siblings, 0 replies; 12+ messages in thread
From: Thomas Lamprecht @ 2026-11-19 22:28 UTC (permalink / raw)
To: pve-devel, Dominik Rusovac
On Thu, 19 Nov 2026 10:59:21 +0100, Dominik Rusovac wrote:
> this picks up Thomas' v3 [0], rebased onto current master and reworked
> in a few places; thx for the nice work Thomas! I added `Originally-by`
> trailers
>
> node maintenance mode can currently only be enabled/disabled with the HA
> manager's CRM command in the CLI. this series adds an API endpoint and a
> 'Node Maintenance' menu in the HA status panel for it
>
> [...]
Applied, thanks!
I folded some changes into the patches, and added two follow-ups:
ha-manager:
- the shared check returns the log level and reason instead of taking a
log callback, so that the CRM and the CLI each phrase the message; the
'not online' message uses the common format now too
- cli: requesting maintenance mode again for a node it is already
requested for succeeds, instead of failing with the CRM's "already in
maintenance state" log message. The early check only runs while the
recorded state is current, i.e. with an active CRM and no other queued
command such as arm-ha, otherwise the CRM decides
- status API: parse the queued commands in PVE::HA::Config
- follow-up: the CRM now records an enable request for a node that is in
maintenance without one, e.g. while it shuts down with the 'migrate'
shutdown policy, so that it stays in maintenance once back online.
This also keeps an enable queued right after a disable
manager:
- the endpoint requires Sys.Modify on / instead of /nodes/{node}, as
maintenance moves HA resources to other nodes; the UI checks the same
- ui: Enable is also disabled while HA is still disarming. The spinner
tracks pending requests in a Map, as a node named 'constructor' was
always pending, keeps them on failed loads and compares them to the
start of the load that finished
docs:
- mention that requests for nodes that are not online are rejected,
except during a 'migrate' policy shutdown, which a follow-up documents
[pve-docs] applied to pve-docs.git (master)
[1/1] ha-manager: maintenance mode: mention web ui menu
https://git.proxmox.com/?p=pve-docs.git;a=commitdiff;h=744af141b66235061a729ba226bd2b0644f6cf25
[pve-ha-manager] applied to pve-ha-manager.git (master)
[1/6] test: add test for ignoring maintenance mode request
https://git.proxmox.com/?p=pve-ha-manager.git;a=commitdiff;h=122e20ce41c19bd62506413b91b39b0e13e4f5fe
[2/6] manager: standardize log messages for maintenance mode request
https://git.proxmox.com/?p=pve-ha-manager.git;a=commitdiff;h=9aeb853a02b80596c346501d76e6fec1f6b11c12
[3/6] manager: factor out check for maintenance mode request
https://git.proxmox.com/?p=pve-ha-manager.git;a=commitdiff;h=b7787287ec1565ed517d7f61f1b175062dc4d808
[4/6] cli: reject maintenance mode requests rejected by crm
https://git.proxmox.com/?p=pve-ha-manager.git;a=commitdiff;h=669933296ffd2854b96e7cf64014d62aaf737a11
[5/6] api: status: report maintenance mode request for lrm entries
https://git.proxmox.com/?p=pve-ha-manager.git;a=commitdiff;h=a885666e11be7365209cd189b9a253e044aa5091
[6/6] api: status: report queued maintenance mode request for lrm entries
https://git.proxmox.com/?p=pve-ha-manager.git;a=commitdiff;h=ef9c125a0277a80289e045a917eaa342fb17199f
[pve-manager] applied to pve-manager.git (master)
[1/3] fix #6144: api: nodes: add endpoint to set maintenance mode
https://git.proxmox.com/?p=pve-manager.git;a=commitdiff;h=b65fde432d9a6a5d529b54a10e8bd6126bf3e978
[2/3] fix #6144: ui: ha: status: allow toggling maintenance mode
https://git.proxmox.com/?p=pve-manager.git;a=commitdiff;h=96a8910070db987bb063f21f834aa97298e0d215
[3/3] ui: ha: status: show spinner while maintenance mode request is queued
https://git.proxmox.com/?p=pve-manager.git;a=commitdiff;h=62c6654ee8eaeddec412dd8b4d053c946cb9b29b
^ permalink raw reply [flat|nested] 12+ messages in thread