[
https://issues.apache.org/jira/browse/BEAM-5708?focusedWorklogId=153897&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-153897
]
ASF GitHub Bot logged work on BEAM-5708:
----------------------------------------
Author: ASF GitHub Bot
Created on: 12/Oct/18 14:27
Start Date: 12/Oct/18 14:27
Worklog Time Spent: 10m
Work Description: tweise commented on a change in pull request #6638:
[BEAM-5708] Cache environment in portable flink runner
URL: https://github.com/apache/beam/pull/6638#discussion_r224803034
##########
File path:
runners/flink/src/main/java/org/apache/beam/runners/flink/translation/functions/ReferenceCountingFlinkExecutableStageContextFactory.java
##########
@@ -104,9 +108,21 @@ private void scheduleRelease(JobInfo jobInfo) {
WrappedContext wrapper = getCache().get(jobInfo.jobId());
Preconditions.checkState(
wrapper != null, "Releasing context for unknown job: " +
jobInfo.jobId());
- // Do not release this asynchronously, as the releasing could fail due to
the classloader not being
- // available anymore after the tasks have been removed from the execution
engine.
- release(wrapper);
+
+ PipelineOptions pipelineOptions =
+ PipelineOptionsTranslation.fromProto(jobInfo.pipelineOptions());
+ int environmentCacheTTLMillis =
+
pipelineOptions.as(PortablePipelineOptions.class).getEnvironmentCacheMillis();
+ if (environmentCacheTTLMillis > 0) {
+ // Schedule task to clean the container later.
+ // Ensure that this class is loaded in the parent Flink classloader.
+ getExecutor()
+ .schedule(() -> release(wrapper), environmentCacheTTLMillis,
TimeUnit.MILLISECONDS);
Review comment:
Something like this (disclaimer: not tested):
```
if (environmentCacheTTLMillis > 0 && this.getClass().getClassLoader() ==
ExecutionEnvironment.class.getClassLoader())
```
For execution in the job server, class loader will be same (applies for
Jenkins). On the remote Flink cluster (by default), the user class loader will
be different and will be removed.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
Issue Time Tracking
-------------------
Worklog Id: (was: 153897)
Time Spent: 1h 20m (was: 1h 10m)
> Support caching of SDKHarness environments in flink
> ---------------------------------------------------
>
> Key: BEAM-5708
> URL: https://issues.apache.org/jira/browse/BEAM-5708
> Project: Beam
> Issue Type: Improvement
> Components: runner-flink
> Reporter: Ankur Goenka
> Assignee: Ankur Goenka
> Priority: Major
> Time Spent: 1h 20m
> Remaining Estimate: 0h
>
> Cache and reuse environment to improve performance.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
