flink-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jürgen Thomann <juergen.thom...@innogames.com>
Subject Changing timeout for cancel command
Date Wed, 12 Apr 2017 11:04:15 GMT

We currently get the following exception if we cancel a job which writes 
to Hadoop:
ERROR org.apache.flink.streaming.connectors.fs.bucketing.BucketingSink  
- Error while trying to hflushOrSync! java.io.InterruptedIOException: 
Interrupted while waiting for data to be acknowledged by pipeline

This causes problem if we cancel a job with creating a savepoint and 
resubmitting the job because the file is sometimes at the end smaller 
than the file size specified in the valid-length file.

Is there a way to increase the time out during cancel to give the flush 
a bit more time? We currently lose events if this happens.


View raw message