Not able to process into Elasticsearch Found

[douwevdijk]

Hi,

We just migrated to a cloud solution using Found. However, when trying to process data into the cluster we are constantly getting the same error:

java.lang.StringIndexOutOfBoundsException: String index out of range: -1

We did not have this issue on our local machine ( using the exact same statement ) so maybe this has something to do with Found?

More details:

Script using: PIG
Script statement: STORE bla INTO 'blabla' USING org.elasticsearch.hadoop.pig.EsStorage('es.nodes=http://xxx.eu-west-1.aws.found.io:9200');

I also tried without using port number, no luck.

[ebuildy]

I got a similar error:

java.lang.StringIndexOutOfBoundsException: String index out of range: -1
    at java.lang.String.substring(String.java:1967)
    at org.elasticsearch.hadoop.rest.RestClient.discoverNodes(RestClient.java:110)
    at org.elasticsearch.hadoop.rest.InitializationUtils.discoverNodesIfNeeded(InitializationUtils.java:58)
    at org.elasticsearch.hadoop.rest.RestService.findPartitions(RestService.java:227)
    at org.elasticsearch.hadoop.mr.EsInputFormat.getSplits(EsInputFormat.java:457)
    at org.elasticsearch.hadoop.mr.EsInputFormat.getSplits(EsInputFormat.java:438)
    at org.apache.spark.rdd.NewHadoopRDD.getPartitions(NewHadoopRDD.scala:115)
    at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:239)
    at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:237)

Using ES / Spark / Python, running a simple:

rdd = sc.newAPIHadoopRDD("org.elasticsearch.hadoop.mr.EsInputFormat","org.apache.hadoop.io.NullWritable", "org.elasticsearch.hadoop.mr.LinkedMapWritable",conf={"es.resource" :"events/events"})

It looks like an issue with elasticSearch server version?

[ebuildy]

Are you using the v2.1.0?

Here https://github.com/elastic/elasticsearch-hadoop/blob/v2.1.0/mr/src/main/java/org/elasticsearch/hadoop/rest/RestClient.java#L110 this cannot work, since ES servers gives:

http_address: "172.17.0.2:9200"

[costin]

Guys, what version of ES-Hadoop are you using? 2.1.2 is not compatible with Elasticsearch 2.0 - only ES-Hadoop 2.2 (currently out at beta1).

[ebuildy]

Yes got it! Was this reason.
Le 15 déc. 2015 7:48 PM, "Costin Leau" notifications@github.com a écrit :

Guys, what version of ES-Hadoop are you using? 2.1.2 is not compatible
with Elasticsearch 2.0 - only ES-Hadoop 2.2 (currently out at beta1).

—
Reply to this email directly or view it on GitHub
#622 (comment)
.

[douwevdijk]

I tried it, still no luck. Does anyone know which version Found is using?

[costin]

@douwevdijk Can you be a bit more specific about what does not work?
Also, have you read the section on cloud/restricted environments?

Without enough information, it's hard to offer any solution...

[douwevdijk]

Hi Costin,

Yes I have read this. I have tried processing with just the most minimal security settings and I checked if any IP was blocked.

But: I am constantly getting the same error and cannot explain it: java.lang.StringIndexOutOfBoundsException: String index out of range: -1

Theoretically this should mean there is something wrong with my code, but I also tried with just processing 1 field. See below my code:

register /usr/lib/pig/lib/elasticsearch-hadoop-pig-2.2.0-beta1.jar
register /usr/lib/pig/lib/piggybank.jar;

d = LOAD 'leads.csv' using org.apache.pig.piggybank.storage.CSVExcelStorage(',', 'NO_MULTILINE') AS (lead_id:chararray, cmp:chararray, status: chararray);

STORE d INTO 'chin/leads' USING org.elasticsearch.hadoop.pig.EsStorage('es.nodes=http://XXX.eu-west-1.aws.found.io:9200', 'es.mapping.pig.tuple.use.field.names:true');

FYI, I tried without setting port to 9200 as this is supposed to be default but got a node not available error.

[costin]

Please turn on logging and post them as gist. The error indicates version 2.2.0 is not used but rather an older version (2.1.0) potentially which is probably still available in the classpath.

[douwevdijk]

Hi Costin,

See here: https://gist.github.com/douwevdijk/1a76b4e061d65e6e9b14

I tried to enable logging but I am not sure this gist includes the lines you need.

[costin]

Right, there's no logging only the exception. By the way, you should see in the job start-up the ES-Hadoop version used.

One more thing, can you try the latest dev snapshot?

[douwevdijk]

Hi,

Indeed PIG was using the 2.1 jar file. So I have succesfully imported the latest .jar ( again). But now I have to dealer with a connection timeout ( didnt had that with the previous .jar )...

[douwevdijk]

OK, it works now! After enabling es.nodes.wan.only. Thanks for the support.