[ 
https://issues.apache.org/jira/browse/HDDS-14843?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Wei-Chiu Chuang updated HDDS-14843:
-----------------------------------
    Release Note: 
OM user/group blacklist (HDDS-14843) — The Ozone Manager can deny access for 
configured users and groups cluster-wide: a full blacklist blocks all 
operations, and a read blacklist blocks only reads. Blacklisted principals are 
denied even if they are also listed as admins. Lists are reconfigurable at 
runtime for faster incident response. Blacklist updates apply per OM (not via 
Ratis), so OMs can disagree until settings match; favor the read blacklist 
first when tightening access.

Configuration: Four new comma-separated list properties (empty by default): 
ozone.blacklist.users and ozone.blacklist.groups for full denial; 
ozone.read.blacklist.users and ozone.read.blacklist.groups for read-only 
denial. They are documented in ozone-default.xml and exposed as OzoneConfigKeys 
constants for tools and tests.

> Support cluster-wide blacklist on OM
> ------------------------------------
>
>                 Key: HDDS-14843
>                 URL: https://issues.apache.org/jira/browse/HDDS-14843
>             Project: Apache Ozone
>          Issue Type: New Feature
>            Reporter: Ivan Andika
>            Assignee: Ivan Andika
>            Priority: Major
>              Labels: pull-request-available
>             Fix For: 2.2.0
>
>
> We have supported an internal user / group blacklist mechanisms to block some 
> anomalous users from destabilizing the cluster for a while. The idea is 
> similar to Ozone admin, but instead of allowing all access, blacklist denies 
> all operations. The configuration is made reconfigurable to allow quick 
> reaction after detection.
> We should prefer to set readonly blacklist first before setting blacklist 
> since (re)configuration is local to each OM (and not applied using Ratis), 
> which might cause state divergence. Although some OM request already push the 
> permission check to preExecute (not validateAndUpdateCache), there are still 
> some requests (mostly multi-keys OM requests) still check in 
> validateAndUpdateCache



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to