[
https://issues.apache.org/jira/browse/ZOOKEEPER-4681?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
krystal he updated ZOOKEEPER-4681:
----------------------------------
Priority: Critical (was: Major)
> uncommitted request of setting data executed
> --------------------------------------------
>
> Key: ZOOKEEPER-4681
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-4681
> Project: ZooKeeper
> Issue Type: Bug
> Components: quorum
> Affects Versions: 3.5.8
> Reporter: krystal he
> Priority: Critical
>
> {color:#172b4d}When the cluster is restarted, the uncommitted requests may be
> silently persisted in the replica.{color}
> {color:#172b4d}Here is a scenario:{color}
> {color:#172b4d}Initial condition:
> Start the ensemble with three nodes A, B and C with C being the leader{color}
> The current epoch is 1
> For simplicity of the example, let’s say zxid is a two digit number, with
> epoch being the first digit
> Create two znodes ‘key0’ and ‘key1’ whose value is ‘0’ and ‘1’, respectively
> The zxid is 12 – 11 for creating key0 and 12 for creating key1. (For
> simplicity of the example, the zxid gets increased only by transactions
> directly changing the data of znodes.)
> All the nodes have seen the change 12 and have persistently logged it
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
