No meaningful logs from bulk write operation failure

(rajesh kumar kalluri) #1

I am trying to index some geo_point data using apache spark, the operation fails with out giving me any information on what is wrong and how I can fix it. When I index the field as string everything checks out. Any clues on how to configure a bulk error handler to print meaningful message and move on. Currently spark crashes after some time.

I also have ignore_malformed set to true on elastic side so am guessing this is happening on spark side of things and not elastic.

"settings": {
"number_of_replicas" : 0,
"index.mapping.ignore_malformed": "true"

Here is a fun stacktrace:

ERROR Executor: Exception in task 7.1 in stage 0.0 (TID 13)
org.elasticsearch.hadoop.EsHadoopException: Could not write all entries for bulk operation [1/100]. Error sample (first [5] error messages):
failed to parse
Bailing out...
at org.elasticsearch.spark.rdd.EsRDDWriter.write(EsRDDWriter.scala:67)
at org.elasticsearch.spark.sql.EsSparkSQL$$anonfun$saveToEs$1.apply(EsSparkSQL.scala:101)
at org.elasticsearch.spark.sql.EsSparkSQL$$anonfun$saveToEs$1.apply(EsSparkSQL.scala:101)
at org.apache.spark.scheduler.ResultTask.runTask(ResultTask.scala:87)
at org.apache.spark.executor.Executor$
at java.util.concurrent.ThreadPoolExecutor.runWorker(
at java.util.concurrent.ThreadPoolExecutor$

(James Baiera) #2

I agree, that is indeed a totally useless error message. I've opened in response.

(system) #3

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.