Disable NFSv4 (server) on Debian, allow NFSv3

最新推荐文章于 2025-05-27 20:46:43 发布

bxklili

最新推荐文章于 2025-05-27 20:46:43 发布

阅读量1k

点赞数

分类专栏：存储技术

存储技术专栏收录该内容

9 篇文章

订阅专栏

本文介绍如何将Debian 8系统的NFS服务器限制为仅支持NFSv3版本。默认情况下，NFS服务器允许NFSv3和NFSv4版本的挂载。文中提供了解决方案，包括修改配置文件中的RPCNFSDCOUNT变量，并重启NFS服务进行验证。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

问题：

How can the NFS server on a Debian 8 system be limited to NFSv3?

By default, shares can be mounted with both vers=3 and vers=4.

/etc/default/nfs-kernel-server:

# To disable NFSv4 on the server, specify '--no-nfs-version 4' here
#RPCMOUNTDOPTS="--manage-gids"
RPCMOUNTDOPTS="--manage-gids --no-nfs-version 4"

This option does not seem to have any effect (rpcinfo still shows nfs accepting version 4).

解决：

2 down vote accepted

Turns out modifying the RPCMOUNTDOPTS variable as described in /etc/default/nfs-kernel-serverdoes not work and there's a bug report for that: #738063

This variable is used in the rpc.mountd call:

# systemctl status nfs-kernel-server
● nfs-kernel-server.service - LSB: Kernel NFS server support
   Loaded: loaded (/etc/init.d/nfs-kernel-server)
   Active: active (running) since Sun 2016-06-12 19:46:01 CEST; 6s ago
  Process: 15110 ExecStop=/etc/init.d/nfs-kernel-server stop (code=exited, status=0/SUCCESS)
  Process: 15119 ExecStart=/etc/init.d/nfs-kernel-server start (code=exited, status=0/SUCCESS)
   CGroup: /system.slice/nfs-kernel-server.service
           └─15167 /usr/sbin/rpc.mountd --manage-gids --port 2048 --no-nfs-version 4

However, clients are still able to mount using -o vers=4.

Instead, this option must be passed to rpc.nfsd. Looking at the init script /etc/init.d/nfs-kernel-server, it seems like the RPCNFSDCOUNT variable is the only variable that's passed to rpc.nfsd. It's not intended for that purpose, but it works and it seems to be the only option short of editing the init script.

Solution:

In /etc/default/nfs-kernel-server, add the --no-nfs-version 4 option to RPCNFSDCOUNT instead of RPCMOUNTDOPTS:

# Number of servers to start up
#RPCNFSDCOUNT=8
RPCNFSDCOUNT="8 --no-nfs-version 4"

Restart the NFS service:

# systemctl restart nfs-kernel-server

Test it:

# mount -t nfs -o vers=4 SERVER:/data/public /mnt
mount.nfs: Protocol not supported

Version 3 still works:

# mount -t nfs -o vers=3 SERVER:/data/public /mnt

2 down vote accepted

Turns out modifying the RPCMOUNTDOPTS variable as described in /etc/default/nfs-kernel-serverdoes not work and there's a bug report for that: #738063

This variable is used in the rpc.mountd call:

# systemctl status nfs-kernel-server
● nfs-kernel-server.service - LSB: Kernel NFS server support
   Loaded: loaded (/etc/init.d/nfs-kernel-server)
   Active: active (running) since Sun 2016-06-12 19:46:01 CEST; 6s ago
  Process: 15110 ExecStop=/etc/init.d/nfs-kernel-server stop (code=exited, status=0/SUCCESS)
  Process: 15119 ExecStart=/etc/init.d/nfs-kernel-server start (code=exited, status=0/SUCCESS)
   CGroup: /system.slice/nfs-kernel-server.service
           └─15167 /usr/sbin/rpc.mountd --manage-gids --port 2048 --no-nfs-version 4

However, clients are still able to mount using -o vers=4.

Instead, this option must be passed to rpc.nfsd. Looking at the init script /etc/init.d/nfs-kernel-server, it seems like the RPCNFSDCOUNT variable is the only variable that's passed to rpc.nfsd. It's not intended for that purpose, but it works and it seems to be the only option short of editing the init script.

Solution:

In /etc/default/nfs-kernel-server, add the --no-nfs-version 4 option to RPCNFSDCOUNT instead of RPCMOUNTDOPTS:

# Number of servers to start up
#RPCNFSDCOUNT=8
RPCNFSDCOUNT="8 --no-nfs-version 4"

Restart the NFS service:

# systemctl restart nfs-kernel-server

Test it:

# mount -t nfs -o vers=4 SERVER:/data/public /mnt
mount.nfs: Protocol not supported

Version 3 still works:

# mount -t nfs -o vers=3 SERVER:/data/public /mnt