[PATCH docs 0/4] ha-manager: add disarming HA section

[PATCH docs 0/4] ha-manager: add disarming HA section

[Maximiliano Sandoval]

Adds documentation around the new

ha-manager crm-command disarm-ha <RESOURCE_MODE>
ha-manager crm-command arm-ha

commands.

Maximiliano Sandoval (4):
  ha-manager: fix enabled typo
  ha-manager: rename node maintenance section
  ha-manager: add disarming HA section
  corosync config: add xref to disarming HA section

 ha-manager.adoc | 37 ++++++++++++++++++++++++++++++++++---
 pvecm.adoc      |  6 ++++++
 2 files changed, 40 insertions(+), 3 deletions(-)

-- 
2.47.3

[PATCH docs 1/4] ha-manager: fix enabled typo

[Maximiliano Sandoval]

Signed-off-by: Maximiliano Sandoval <m.sandoval@proxmox.com>
---
 ha-manager.adoc | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/ha-manager.adoc b/ha-manager.adoc
index 4c318fb..859721d 100644
--- a/ha-manager.adoc
+++ b/ha-manager.adoc
@@ -1116,7 +1116,7 @@ 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
+Currently you can enable or disable the maintenance mode using the ha-manager
 CLI tool.
 
 .Enabling maintenance mode for a node
-- 
2.47.3

[PATCH docs 2/4] ha-manager: rename node maintenance section

[Maximiliano Sandoval]

We will add a new section documenting how to disarm HA named
"cluster-wide maintenance", in order to make them semantically easier to
understand, we rename the existing "node maintenance" section.

Signed-off-by: Maximiliano Sandoval <m.sandoval@proxmox.com>
---
 ha-manager.adoc | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/ha-manager.adoc b/ha-manager.adoc
index 859721d..68fb7d1 100644
--- a/ha-manager.adoc
+++ b/ha-manager.adoc
@@ -1087,8 +1087,8 @@ case, may result in a reset triggered by the watchdog.
 
 
 [[ha_manager_node_maintenance]]
-Node Maintenance
-----------------
+Single Node Maintenance
+-----------------------
 
 Sometimes it is necessary to perform maintenance on a node, such as replacing
 hardware or simply installing a new kernel image. This also applies while the
-- 
2.47.3

[PATCH docs 3/4] ha-manager: add disarming HA section

[Maximiliano Sandoval]

Signed-off-by: Maximiliano Sandoval <m.sandoval@proxmox.com>
---
 ha-manager.adoc | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)

diff --git a/ha-manager.adoc b/ha-manager.adoc
index 68fb7d1..3c97a2d 100644
--- a/ha-manager.adoc
+++ b/ha-manager.adoc
@@ -1086,6 +1086,37 @@ this is not the case the update process can take too long which, in the worst
 case, may result in a reset triggered by the watchdog.
 
 
+[[ha_manager_cluster_maintenance]]
+Cluster-wide Maintenance
+------------------------
+
+If a cluster has HA-resources then one has to be careful during a maintenance
+that involves multiple nodes or when performing configuration changes on
+multiple node, as connectivity issues on the Corosync network can result in one
+or more nodes fencing. One common example is when editing Corosync's
+configuration as the changes will be picked up by all cluster nodes and any
+configuration error can lead to nodes potentially fencing. In this scenario one
+can temporarily disarm HA in order to release the watchdogs in the cluster.
+
+You can disarm and arm HA using the ha-manager CLI tool.
+
+.Disarming HA
+----
+# ha-manager crm-command disarm-ha <RESOURCE_MODE>
+----
+
+.Arming HA
+----
+# ha-manager crm-command arm-ha
+----
+
+There are two possible resource modes for disarming HA:
+
+- `freeze`: new commands and state changes are not applied.
+- `ignore`: resources are removed from HA tracking and can be managed as if they
+  were not HA managed.
+
+
 [[ha_manager_node_maintenance]]
 Single Node Maintenance
 -----------------------
-- 
2.47.3

[PATCH docs 4/4] corosync config: add xref to disarming HA section

[Maximiliano Sandoval]

Signed-off-by: Maximiliano Sandoval <m.sandoval@proxmox.com>
---
 pvecm.adoc | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/pvecm.adoc b/pvecm.adoc
index a48b18c..e4330ae 100644
--- a/pvecm.adoc
+++ b/pvecm.adoc
@@ -1377,6 +1377,12 @@ corosync service via:
 systemctl restart corosync
 ----
 
+Note that restarting the Corosync service on a cluster with HA resources might
+lead to the node fencing if the service does not come back up properly. This can
+happen, for example, if the new version of the configuration has a typo or
+errors in general. xref:ha_manager_cluster_maintenance[Cluster-wide Maintenance]
+for more details.
+
 On errors, check the troubleshooting section below.
 
 Changing the Token Coefficient
-- 
2.47.3

Re: [PATCH docs 0/4] ha-manager: add disarming HA section

[Thomas Lamprecht]

Am 30.03.26 um 15:40 schrieb Maximiliano Sandoval:
> Adds documentation around the new
> 
> ha-manager crm-command disarm-ha <RESOURCE_MODE>
> ha-manager crm-command arm-ha
> 

Thanks, but there was already docs for the [v1], I now pushed a slightly
updated variant of that. The differences are mostly adapting to the change
that resources are not removed from HA manager status but rather their
tracking is just suspended, i.e. mostly s/remove/suspend/;

Please send any improvements on top of that.

[v1]: https://lore.proxmox.com/all/20260310155216.2086316-1-t.lamprecht@proxmox.com/