Access Control Lists (ACLs) on HDFS help in addressing access-related issues better than Permission Bits.
Drawbacks of Permission Bits
Before the implementation of ACls, the HDFS permission model was equivalent to traditional UNIX Permission Bits. In this model, permissions for each file or directory are managed by a set of three distinct user classes: Owner, Group, and Others. There are three permissions for each user class: Read, Write, and Execute. Thus, for any file system object, its permissions can be encoded in 3*3=9 bits. When a user attempts to access a file system object, HDFS enforces permissions according to the most specific user class applicable to that user. If the user is the owner, HDFS checks the Owner class permissions. If the user is not the owner, but is a member of the file system object’s group, HDFS checks the Group class permissions. Otherwise, HDFS checks the Others class permissions.
This model can sufficiently address a large number of security requirements. For example, consider a sales department that would like a single user -- Bruce, the department manager -- to control all modifications to sales data. Other members of the sales department need to view the data, but must not be allowed to modify it. Everyone else in the company (outside of the sales department) must not be allowed to view the data. This requirement can be implemented by running chmod 640 on the file, with the following outcome:
-rw-r-----1 brucesales22K Nov 18 10:55 sales-data
Only Bruce can modify the file, only members of the sales group can read the file, and no one else can access the file in any way.
Suppose new requirements arise as a result of which users Bruce, Diana, and Clark are allowed to make modifications. Permission Bits cannot address this requirement, because there can be only one owner and one group, and the group is already used to implement the read-only requirement for the sales team. A typical workaround is to set the file owner to a synthetic user account, such as "salesmgr," and allow Bruce, Diana, and Clark to use the "salesmgr" account through sudo or similar impersonation mechanisms. The drawback with this workaround is that it forces complexity on end-users, requiring them to use different accounts for different actions.
Consider another example where the sales staff and all the executives require access to the sales data. This is another requirement that Permission Bits cannot address, because there is only one group, and it is already used by sales. A typical workaround is to set the file’s group to a new synthetic group, such as "salesandexecs," and add all users of "sales" and all users of "execs" to that group. The drawback with this workaround is that it requires administrators to create and manage additional users and groups.
The preceding examples indicate that Permission Bits cannot address permission requirements that differ from the natural organizational hierarchy of users and groups. ACLs enable you to address these requirements more naturally by allowing multiple users and multiple groups to have different sets of permissions.
Examples of using ACLs for addressing access permission requirements
The following examples explain how you can use ACLs and address different requirements related to access permissions:
Using a Default ACL for Automatic Application to New Children
In addition to an ACL that is enforced during permission checks, a default ACL is also available. A default ACL can only be applied to a directory -- not to a file. Default ACLs have no direct effect on permission checks for existing child files and directories, but instead define the ACL that new child files and directories will receive when they are created.
Suppose we have a "monthly-sales-data" directory that is further subdivided into separate directories for each month. We will set a default ACL to guarantee that members of the "execs" group automatically get access to new subdirectories as they get created each month.
Set a default ACL on the parent directory:
> hdfs dfs -setfacl -m default:group:execs:r-x /monthly-sales-data
> hdfs dfs -mkdir /monthly-sales-data/JAN > hdfs dfs -mkdir /monthly-sales-data/FEB
Verify that HDFS has automatically applied the default ACL to the subdirectories:
> hdfs dfs -getfacl -R /monthly-sales-data # file: /monthly-sales-data # owner: bruce # group: sales user::rwx group::r-x other::--- default:user::rwx default:group::r-x default:group:execs:r-x default:mask::r-x default:other::--- # file: /monthly-sales-data/FEB # owner: bruce # group: sales user::rwx group::r-x group:execs:r-x mask::r-x other::--- default:user::rwx default:group::r-x default:group:execs:r-x default:mask::r-x default:other::--- # file: /monthly-sales-data/JAN # owner: bruce # group: sales user::rwx group::r-x group:execs:r-x mask::r-x other::--- default:user::rwx default:group::r-x default:group:execs:r-x default:mask::r-x default:other::---
Blocking Access to a Sub-Tree for a Specific User
Suppose there is a need to immediately block access to an entire sub-tree for a specific user. Applying a named user ACL entry to the root of that sub-tree is the fastest way to accomplish this without accidentally revoking permissions for other users.
Add an ACL entry to block user Diana's access to "monthly-sales-data":
> hdfs dfs -setfacl -m user:diana:--- /monthly-sales-data
Run getfacl to check the results:
> hdfs dfs -getfacl /monthly-sales-data # file: /monthly-sales-data # owner: bruce # group: sales user::rwx user:diana:--- group::r-x mask::r-x other::--- default:user::rwx default:group::r-x default:group:execs:r-x default:mask::r-x default:other::---
It is important to keep in mind the order of evaluation for ACL entries when a user attempts to access a file system object:
If the user is the file owner, the Owner Permission Bits are enforced.
Else, if the user has a named user ACL entry, those permissions are enforced.
Else, if the user is a member of the file’s group or any named group in an ACL entry, then the union of permissions for all matching entries are enforced. (The user may be a member of multiple groups.)
If none of the above are applicable, the Other Permission Bits are enforced.
In this example, the named user ACL entry accomplished our goal because the user is not the file owner and the named user entry takes precedence over all other entries.