[ 
https://issues.apache.org/jira/browse/KAFKA-3890?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Henry Cai updated KAFKA-3890:
-----------------------------
    Description: 
Currently the task assignment in TaskAssignor is not deterministic.  During 
cluster restart or rolling restart, even though the participating worker nodes 
are the same, but the TaskAssignor is not able to maintain a deterministic 
mapping, so about 20% partitions will be reassigned which would cause state 
repopulation on cluster restart time.

When the participating worker nodes are not changed, we really just want to 
keep the old task assignment.


  was:
Today most of the rocksDB configs are hard written inside {{RocksDBStore}}, or 
the default values are directly used. We need to make them configurable for 
advanced users. For example, some default values may not work perfectly for 
some scenarios: 
https://github.com/HenryCaiHaiying/kafka/commit/ccc4e25b110cd33eea47b40a2f6bf17ba0924576
 

One way of doing that is to introduce a "RocksDBStoreConfigs" objects similar 
to "StreamsConfig", which defines all related rocksDB options configs, that can 
be passed as key-value pairs to "StreamsConfig".


> Kafka Streams: task assignment is not maintained on cluster restart or 
> rolling restart
> --------------------------------------------------------------------------------------
>
>                 Key: KAFKA-3890
>                 URL: https://issues.apache.org/jira/browse/KAFKA-3890
>             Project: Kafka
>          Issue Type: Bug
>          Components: streams
>            Reporter: Henry Cai
>            Assignee: Henry Cai
>              Labels: api, newbie
>
> Currently the task assignment in TaskAssignor is not deterministic.  During 
> cluster restart or rolling restart, even though the participating worker 
> nodes are the same, but the TaskAssignor is not able to maintain a 
> deterministic mapping, so about 20% partitions will be reassigned which would 
> cause state repopulation on cluster restart time.
> When the participating worker nodes are not changed, we really just want to 
> keep the old task assignment.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to