lengyuexuexuan commented on code in PR #101:
URL:
https://github.com/apache/incubator-pegasus-website/pull/101#discussion_r2035293794
##########
_docs/en/administration/rebalance.md:
##########
@@ -2,4 +2,380 @@
permalink: administration/rebalance
---
-TRANSLATING
+This document mainly introduces the concepts, usage, and design of rebalance
in Pegasus.
+
+## Concept Section
+In Pegasus, rebalance mainly includes the following aspects:
+1. If a certain partition replica does not meet the requirement of one primary
and two secondaries, a node needs to be selected to complete the missing
replica. This process in Pegasus is called `cure`.
+2. After all partitions meet the requirement of one primary and two
secondaries, the number of replicas on each replica server in the cluster is
adjusted to try to keep the number of replicas served by each machine at a
similar level. This process in Pegasus is called `balance`.
+3. If a replica server has multiple disks mounted and provided to Pegasus for
use through the configuration file `data_dirs`, the replica server should try
to keep the number of replicas on each disk at a similar level.
+
+Based on these points, Pegasus has introduced some concepts to conveniently
describe these situations:
+1. The Health Status of Partition
+ Pegasus has defined several health statuses for partitions:
+ * 【fully healthy】: Healthy, fully meeting the requirement of one primary
and two secondaries.
+ * 【unreadable】: The partition is unreadable. It means that the partition
lacks a primary, but there is one or two secondaries.
+ * 【readable but unwritable】: The partition is readable but not writable. It
means that only one primary remains, and both secondary replicas are lost.
+ * 【readable and writable but unhealthy】: The partition is both readable and
writable, but still not healthy. It means that one secondary is missing among
the three replicas.
+ * 【dead】: All replicas of the partition are unavailable, also known as the
DDD state.
+{:class="img-responsive"}
+
+ When checking the status of the cluster, tables, and partitions through the
Pegasus shell, you will often see the overall statistics or individual
descriptions of the health conditions of the partitions. For example, by using
the ls -d command, you can see the number of partitions in different health
conditions for each table, including the following:
Review Comment:
Thanks, all suggestions committed.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
