[
https://issues.apache.org/jira/browse/KAFKA-3184?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16736113#comment-16736113
]
Guozhang Wang commented on KAFKA-3184:
--------------------------------------
[~Yohan123] I think you are right: currently `persistent()` is used to
determine whether we should record checkpoint offsets, since for non-persistent
stores there are no data flushed to persistent storage and therefore we can
only restore from beginning every time upon resuming. With this JIRA in-memory
stores would be "persisted" to storage as well, the only difference is that it
is not written until that "snapshotting" point is reached (e.g. after N
commits). And hence in this case persistent() flag is not that useful anyways
and may consider deprecated.
> Add Checkpoint for In-memory State Store
> ----------------------------------------
>
> Key: KAFKA-3184
> URL: https://issues.apache.org/jira/browse/KAFKA-3184
> Project: Kafka
> Issue Type: Improvement
> Components: streams
> Reporter: Guozhang Wang
> Priority: Major
> Labels: user-experience
>
> Currently Kafka Streams does not make a checkpoint of the persistent state
> store upon committing, which would be expensive since it is "stopping the
> world" and write on disks: for example, RocksDB would require you to copy the
> file directory to make a copy naively.
> However, for in-memory stores checkpointing maybe doable in an asynchronous
> manner hence it can be done quickly. And the benefit of having intermediate
> checkpoint is to avoid restoring from scratch if standby tasks are not
> present.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
