How to configure an all_squash equivalent on a NetApp NFS export
Applies to
- Data ONTAP 8.2 7-Mode
- Data ONTAP 8.1 7-Mode
- Data ONTAP 8 7-Mode
- Data ONTAP 7 and earlier
Description
This article describes on how to configure an all_squash equivalent on a NetApp NFS export.
Sometimes users prefer to use the all_squash
option on a Network File System (NFS) export. On the Linux NFS server (not Data ONTAP), the all_squash
option makes the server disregard the incoming NFS User Identifier/ Group Identifier (UID/ GID) and allows the server to be set in the exports instead. When the all_squash
option is used, all clients for a particular export are forced to read and write as the same user.
The all_squash
option is not used on the storage system. However, the sec=none
and anon=(uid)
options can be used. For more information, see the following KB: How to configure no_root_squash on Ontap 7mode and Clustered Data ONTAP.
For more information: TR-4067 NFS Best Practice and Implementation Guide (section: 4.8 Mapping All UIDs to a Single UID (squash_all))
For more information on how to perform this in Clustered Data ONTAP: How to enable the equivalent of all_squash in clustered Data ONTAP
The TSAP discusses that the issue is with all the newer NFS clients (not the storage system) that can be reproduced using two Linux machines. Two Debian boxes were used with the sec=none
command in the export and the -o sec=none
command in the mount and they failed.
Using the sec=none
command with the anon=(uid)
command through NFSv3 fails on the storage system, and the following message is logged on the console:
Client <IP>, is send the NULL reply
When trying to mount, the client will return an error output similar to the following:
When trying to mount, the storage system will return an error output similar to the following: