From e4503d18b491996ef2ea45958bc857be3aba9cd4 Mon Sep 17 00:00:00 2001
From: "copilot-swe-agent[bot]" <198982749+Copilot@users.noreply.github.com>
Date: Sat, 15 Aug 2026 03:38:51 +0000
Subject: [PATCH] feat(chart): add default-nvme-tcp-nr-io-queues setting docs
 for v1.13.0 and v1.12.2

longhorn/longhorn#13706

Co-authored-by: derekbit <12527233+derekbit@users.noreply.github.com>
---
 content/docs/1.12.2/references/settings.md | 7 +++++++
 content/docs/1.13.0/references/settings.md | 7 +++++++
 2 files changed, 14 insertions(+)

diff --git a/content/docs/1.12.2/references/settings.md b/content/docs/1.12.2/references/settings.md
index 38ea010c5..7c44ec140 100644
--- a/content/docs/1.12.2/references/settings.md
+++ b/content/docs/1.12.2/references/settings.md
@@ -54,6 +54,7 @@ weight: 1
   - [Data Engine Log Flags](#data-engine-log-flags)
   - [Replica Rebuilding Bandwidth Limit](#replica-rebuilding-bandwidth-limit)
   - [Manager URL](#manager-url)
+  - [Default NVMe-TCP Number Of IO Queues](#default-nvme-tcp-number-of-io-queues)
   - [Default Ublk Queue Depth](#default-ublk-queue-depth)
   - [Default Ublk Number Of Queue](#default-ublk-number-of-queue)
   - [Node Disk Health Monitoring](#node-disk-health-monitoring)
@@ -651,6 +652,12 @@ This setting is useful when accessing the Longhorn API through Ingress or Gatewa
 
 For more details, see [Manager URL for External API Access](../../important-notes/#manager-url-for-external-api-access).
 
+#### Default NVMe-TCP Number Of IO Queues
+
+> Default: `{"v2":"0"}`
+
+The default number of I/O queues the kernel initiator creates when connecting a volume frontend over NVMe-TCP. This caps the per-volume in-flight commands to the number of I/O queues multiplied by the negotiated queue size (128). This setting applies to volumes using the V2 Data Engine with the block device frontend and takes effect on (re)attach. The value can be overridden per volume using volume.spec.nvmeTcpNrIoQueues. 0 means unspecified (kernel default, one queue per online core).
+
 #### Default Ublk Queue Depth
 
 > Default: `{"v2":"128"}`
diff --git a/content/docs/1.13.0/references/settings.md b/content/docs/1.13.0/references/settings.md
index b25407f32..7ab349ea8 100644
--- a/content/docs/1.13.0/references/settings.md
+++ b/content/docs/1.13.0/references/settings.md
@@ -55,6 +55,7 @@ weight: 1
   - [Data Engine Log Flags](#data-engine-log-flags)
   - [Replica Rebuilding Bandwidth Limit](#replica-rebuilding-bandwidth-limit)
   - [Manager URL](#manager-url)
+  - [Default NVMe-TCP Number Of IO Queues](#default-nvme-tcp-number-of-io-queues)
   - [Default Ublk Queue Depth](#default-ublk-queue-depth)
   - [Default Ublk Number Of Queue](#default-ublk-number-of-queue)
   - [Node Disk Health Monitoring](#node-disk-health-monitoring)
@@ -670,6 +671,12 @@ This setting is useful when accessing the Longhorn API through Ingress or Gatewa
 
 For more details, see [Manager URL for External API Access](../../important-notes/#manager-url-for-external-api-access).
 
+#### Default NVMe-TCP Number Of IO Queues
+
+> Default: `{"v2":"0"}`
+
+This setting defines the default number of I/O queues the kernel initiator creates when connecting a volume frontend over NVMe-TCP. It caps the in-flight commands per volume to the number of I/O queues multiplied by the negotiated queue size (128). This applies to volumes using the V2 Data Engine with the block device frontend, and changes take effect upon volume attachment or reattachment. You can override this value on a per-volume basis using `volume.spec.nvmeTcpNrIoQueues`. 0 means unspecified (kernel default, one queue per online core).
+
 #### Default Ublk Queue Depth
 
 > Default: `{"v2":"128"}`