[
https://issues.apache.org/jira/browse/BEAM-11435?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
David Morávek updated BEAM-11435:
---------------------------------
Fix Version/s: 2.27.0
Resolution: Fixed
Status: Resolved (was: Open)
> Reuse already set timers instead of recreating them in Flink Streaming Runner
> -----------------------------------------------------------------------------
>
> Key: BEAM-11435
> URL: https://issues.apache.org/jira/browse/BEAM-11435
> Project: Beam
> Issue Type: Bug
> Components: runner-flink
> Affects Versions: 2.23.0, 2.24.0, 2.25.0
> Reporter: David Morávek
> Assignee: David Morávek
> Priority: P1
> Fix For: 2.27.0
>
> Time Spent: 50m
> Remaining Estimate: 0h
>
> When we re-set a timer, new timer goes trough following life-cycle:
> 1) Find previously set timer with the same id
> 2) Cancel any previously set timer with the same id
> 3) Set a new timer
> We can skip 2) and 3), if the new timer is the same as previously set timer.
> There should be a great performance impact on GC timers, especially with
> RocksDB backend, where 2) and 3) are expensive operations.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
