Myles Baker commented on AVRO-1419:

I've seen this for Spark jobs using yarn as the resource manager. The job 
aborts due to stage failure caused by  org.apache.avro.AvroRuntimeException: 
java.io.IOException: Invalid sync!

> java.io.IOException: Invalid sync! throw after random number of sync() calls.
> -----------------------------------------------------------------------------
>                 Key: AVRO-1419
>                 URL: https://issues.apache.org/jira/browse/AVRO-1419
>             Project: Avro
>          Issue Type: Bug
>          Components: java
>    Affects Versions: 1.7.5
>            Reporter: Deepak Kumar V
> I have a 340 MB avro data file that contains records sorted and identified by 
> unique id (duplicate records exists). At the beginning of every unique record 
> a synchronization point is created with DataFileWriter.sync(). (I cannot or 
> do not want to save the sync points and i do not want to use 
> SortedKeyValueFile as output format for M/R job)  
> There are at-least 25k synchronization points in a 340 MB file.
> Ex:
> Marker1_RecordA1_RecordA2_RecordA3_Marker2_RecordB1_RecordB2
> As records are sorted and marked, for efficient retrieval, binary search is 
> performed. Most of the times the search is successful, at times the code 
> throws the following exception
> ------
> org.apache.avro.AvroRuntimeException: java.io.IOException: Invalid sync! at 
> org.apache.avro.file.DataFileStream.hasNext(DataFileStream.java:210 
> ------
> I note down the position that was used to invoke fileReader.sync(mid); and 
> catch AvroRuntimeException, close and open the file and sync(mid) i do not 
> see exception. 
> Why should Avro throw exception before and not later ?
> 1.7.5v of library is throwing this error. Raising a major defect, adjust the 
> priority at your convenience. 

This message was sent by Atlassian JIRA

Reply via email to