[
https://issues.apache.org/jira/browse/GOBBLIN-2024?focusedWorklogId=911430&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-911430
]
ASF GitHub Bot logged work on GOBBLIN-2024:
-------------------------------------------
Author: ASF GitHub Bot
Created on: 26/Mar/24 00:12
Start Date: 26/Mar/24 00:12
Worklog Time Spent: 10m
Work Description: AndyJiang99 commented on code in PR #3902:
URL: https://github.com/apache/gobblin/pull/3902#discussion_r1538388499
##########
gobblin-iceberg/src/main/java/org/apache/gobblin/iceberg/publisher/GobblinMCEPublisher.java:
##########
@@ -214,6 +214,7 @@ public static Metrics getMetrics(State state, Path path,
Configuration conf, Nam
return OrcMetrics.fromInputFile(HadoopInputFile.fromPath(path,
conf), MetricsConfig.getDefault(), mapping);
} catch (Exception e) {
//This means the table is not compatible with iceberg, so return a
dummy metric
+ //todo: throw exception here? as it's corrupted file
Review Comment:
These two seem to be contradictory it kind of sounds like. If the table is
not compatible with iceberg, then rather than return a dummy metric (which
indicates the table is not compatible but no underlying issues), we should just
throw the exception due to corrupted file right?
Issue Time Tracking
-------------------
Worklog Id: (was: 911430)
Time Spent: 0.5h (was: 20m)
> Improve the performance of GobblinMCEWriter
> -------------------------------------------
>
> Key: GOBBLIN-2024
> URL: https://issues.apache.org/jira/browse/GOBBLIN-2024
> Project: Apache Gobblin
> Issue Type: Improvement
> Reporter: Zihan Li
> Priority: Major
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> We are seeing slowness in the GobblinMCEWriter, which will cause us huge
> backlog when handling large topic and affect the other topics in the same
> container for GMIP. After benchmark, we identify that we can improve the
> computeSpec part to reduce the duplicate work, also we can add the file size
> information in GMCE to avoid calling HDFS for each file.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
