You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@hudi.apache.org by "suheng.cloud (Jira)" <ji...@apache.org> on 2021/11/01 05:43:00 UTC
[jira] [Created] (HUDI-2659) concurrent compaction problem on flink
sql
suheng.cloud created HUDI-2659:
----------------------------------
Summary: concurrent compaction problem on flink sql
Key: HUDI-2659
URL: https://issues.apache.org/jira/browse/HUDI-2659
Project: Apache Hudi
Issue Type: Bug
Components: Flink Integration
Reporter: suheng.cloud
Attachments: image-2021-11-01-13-14-40-831.png, image-2021-11-01-13-16-28-695.png
Hi, Community:
We continously watch the flink compact task, and found there maybe some issue after the job run 2 days.
The taskmanager log shows that the 2 compact plan executed in sequence, in witch the former commit action delete the basefile(for some duplicated reason?) which was a dependence of the latter one.
I wonder will this cause data lost in final ?
the core flink sink table params are:
{code:java}
'table.type' = 'MERGE_ON_READ','table.type' = 'MERGE_ON_READ', 'write.operation'='upsert', 'read.streaming.enabled' = 'true', 'hive_sync.enable' = 'false', 'write.precombine.field'='ts', 'compaction.trigger.strategy'='num_commits', 'compaction.delta_commits'= '5', 'compaction.tasks'='4', 'compaction.max_memory'='10', 'hoodie.parquet.max.file.size'='20971520', 'hoodie.parquet.small.file.limit'='10485760', 'write.log.max.size'='52428800', 'compaction.target_io'='5120', 'changelog.enabled'='false', 'clean.retain_commits'='20', 'archive.max_commits'='30', 'archive.min_commits'='20'{code}
cc [~danny0405], can you also give some suggestion :)
Thank you all~
!image-2021-11-01-13-14-40-831.png!
!image-2021-11-01-13-16-28-695.png!
--
This message was sent by Atlassian Jira
(v8.3.4#803005)