GitHub user WangTaoTheTonic opened a pull request:
FLINK-6295use LoadingCache instead of WeakHashMap to lower latency
Now in ExecutionGraphHolder, which is used in many handlers, we use a WeakHashMap to cache ExecutionGraph(s), which is only sensitive to garbage collection.
The latency is too high when JVM do GC rarely, which will make status of jobs or its tasks unmatched with the real ones. (WE once observed that the web still shows tasks cancelled/failed, after the actual states of tasks coming back to normal for *30+ mins,* until a gc happened)
LoadingCache is a common used cache implementation from guava lib, we can use its time based eviction to lower latency of status update.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/WangTaoTheTonic/flink
Alternatively you can review and apply these changes as the patch at:
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #3709
Author: WangTaoTheTonic <firstname.lastname@example.org>
use LoadingCache instead of WeakHashMap to lower latency