[
https://issues.apache.org/jira/browse/FLINK-15507?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17483048#comment-17483048
]
Yuan Mei commented on FLINK-15507:
----------------------------------
After offline discussion with [~pnowojski], [~yunta] & [~roman] we decided at
least a few more questions/problems needed to be answered/resolved before local
recovery was enabled by default.
# What extra value can bring to enable local recovery {*}by default{*}?
# It may introduce extra performance (double writing checkpoints) degrade +
extra local disk space when performing full checkpoints with rocksdb + FS
store. I think this may cause unexpected problems for cloud/service providers
when upgrading Flink. We should be careful communicating and discussing this.
# Local recovery should work fine (no significant downside) for incremental
checkpoints with rocksdb. However, incremental is not a default now. So to make
local recovery by default meaningful, {*}we should enable incremental for
rocksdb by default as a pre-requisite{*}.
# Local recovery does not support unaligned checkpoint and changelog state
backend. This is not a blocker, but need to take into account as well.
(backpressure vs recovery).
> Activate local recovery for RocksDB backends by default
> -------------------------------------------------------
>
> Key: FLINK-15507
> URL: https://issues.apache.org/jira/browse/FLINK-15507
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / State Backends
> Reporter: Stephan Ewen
> Assignee: Yuan Mei
> Priority: Major
> Labels: auto-deprioritized-critical, auto-unassigned,
> pull-request-available
>
> For the RocksDB state backend, local recovery has no overhead when
> incremental checkpoints are used.
> It should be activated by default, because it greatly helps with recovery.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
