[
https://issues.apache.org/jira/browse/FLINK-10360?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17267664#comment-17267664
]
Roman Khachatryan commented on FLINK-10360:
-------------------------------------------
[~stevenz3wu],
is it the same issue as FLINK-9465?
If yes should we close it as a duplicate then?
> support timeout in savepoint REST api
> -------------------------------------
>
> Key: FLINK-10360
> URL: https://issues.apache.org/jira/browse/FLINK-10360
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / Checkpointing, Runtime / REST, Runtime / State
> Backends
> Reporter: Steven Zhen Wu
> Priority: Major
>
> right now, savepoint share the same timeout config as checkpoint. With
> incremental checkpoint, we just need to configure the timeout to be
> reasonable for the small delta. With savepoint, it is a full snapshot and
> maybe the checkpoint timeout is too small for savepoint.
> * incremental checkpoint size ranges from 100 GB - 400 GB, and took 1 mins -
> 4 mins
> * full savepoint is ~4.5 TB and took ~25 mins
> It will bring more flexibility if the savepoint REST api supports an
> additional timeout param, and apply it for actual savepoint timeout.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
