Siyao Meng created HDDS-17134:
---------------------------------
Summary: Document OM and SCM HA with more than 3 nodes
Key: HDDS-17134
URL: https://issues.apache.org/jira/browse/HDDS-17134
Project: Apache Ozone
Issue Type: Sub-task
Components: documentation
Reporter: Siyao Meng
Update OM-HA.md, SCM-HA.md (and the zh versions) and the Listener OM design doc:
* A fault-tolerance table: 3 nodes tolerate 1 failure, 5 tolerate 2; an even
count adds no tolerance.
* Correct listener-om.md: "A setup with 2 voting OMs will not work as Ratis
requires an odd number of voting members". Two voters work, they just tolerate
no failure.
* Adding OMs: every client configuration (clients, S3 Gateway, Recon, HttpFS)
must include the new OMs before one can become leader (see the OM client
failover sub-task).
* SCM: the extra datanode report load per added SCM, and the primordial SCM
dependency in secure clusters.
* Expected leadership transfer latency (until the RatisHelper leader hint
sub-task is fixed).
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
