docs: document guest network and disk policies

Author: Vonng

content/blog/design/fixed-ip-two-nics.md

@@ -24,7 +24,12 @@ interfaces.
 | Interface | Addressing | Responsibility |
 | --- | --- | --- |
 | management | QEMU user-mode NAT, DHCP | DNS, default route, outbound internet, and loopback SSH fallback |
-| `private0` | fixed Inventory address | host-to-node, node-to-node, Ansible, Pigsty services, and VIP traffic |
+| private, MAC-matched | fixed Inventory address | host-to-node, node-to-node, Ansible, Pigsty services, and VIP traffic |
+
+Farrow matches both physical interfaces by their deterministic MAC addresses
+and does not rename them. Guest-visible names are whatever the image's network
+stack chooses (commonly `eth0` or `enp0s4`); the MAC and address contract, not
+the display name, determines each interface's role.
 
 The management interface is deliberately ordinary. It lets a new cloud image
 reach package repositories before any application exists, without installing

content/blog/design/fixed-ip-two-nics.zh.md

@@ -21,7 +21,10 @@ PostgreSQL 暴露给宿主，却无法把这种网络身份提供给其它节点
 | 网卡 | 地址方式 | 职责 |
 | --- | --- | --- |
 | 管理网卡 | QEMU User-mode NAT、DHCP | DNS、默认路由、互联网出站与回环 SSH 备用路径 |
-| `private0` | Inventory 固定地址 | 宿主到节点、节点间、Ansible、Pigsty 服务与 VIP 流量 |
+| 私网网卡（按 MAC 匹配） | Inventory 固定地址 | 宿主到节点、节点间、Ansible、Pigsty 服务与 VIP 流量 |
+
+Farrow 使用确定的 MAC 地址匹配两张物理网卡，不再修改它们的名称。Guest 中看到的名字由镜像
+自身的网络栈决定（通常是 `eth0` 或 `enp0s4`）；网卡角色由 MAC 与地址契约决定，而不是显示名称。
 
 管理网卡刻意保持普通：一张新的 Cloud Image 在应用尚未安装前就能访问软件仓库，宿主无需
 安装 NAT 规则或 DHCP 服务。

content/docs/reference/configuration.md

@@ -78,11 +78,14 @@ architecture also needs its matching `qemu-system-*` binary and firmware.
 vm_disks:
   - path: /data
     size: 128
-    fs: xfs
+    fs: auto
     persistent: false
 ```
 
-`path` is the disk identity and mount point. `fs` is `xfs` or `ext4`.
+`path` is the disk identity and mount point. `fs` is `auto`, `xfs`, or `ext4`;
+omitting it retains the released `xfs` default. An explicitly selected `auto`
+disk uses XFS when `mkfs.xfs` is available and otherwise uses ext4. Explicit
+`xfs` and `ext4` never fall back, and an existing filesystem is preserved.
 `persistent: true` survives normal destroy. Use `vm_disks: []` for no extra
 disk.
 

content/docs/reference/configuration.zh.md

@@ -68,12 +68,14 @@ Linux setup 只安装宿主原生模拟器；外来架构还需要对应 `qemu-s
 vm_disks:
   - path: /data
     size: 128
-    fs: xfs
+    fs: auto
     persistent: false
 ```
 
-`path` 同时是磁盘身份与挂载点；`fs` 为 `xfs` 或 `ext4`。`persistent: true` 在普通
-destroy 后保留；`vm_disks: []` 表示不要额外盘。
+`path` 同时是磁盘身份与挂载点；`fs` 为 `auto`、`xfs` 或 `ext4`，省略时保持已发布的
+`xfs` 默认值。显式选择的空白 `auto` 磁盘在 `mkfs.xfs` 可用时优先使用 XFS，否则
+使用 ext4；显式 `xfs`、`ext4` 不会降级，已有文件系统保持不变。`persistent: true`
+在普通 destroy 后保留；`vm_disks: []` 表示不要额外盘。
 
 ## 目录共享