Hi Austin,
The record key is null and it cannot be submitted to Ignite. Please make sure
your key is not null.
Btw, "parse.key" and "key.separator" are ConsoleProducer's properties, not
Worker's.
-Roman
On Monday, November 7, 2016 10:26 PM, austin solomon
<austin.solomon...@gmail.com> wrote:
Hi,
I am using Kafka Streamer module to pump the data from Kafka to Ignite
cache, by following the documentation given in the below link
https://apacheignite.readme.io/docs/kafka-streamer
I could able to sink the data in ignite's cache. However when i tried to
achieve the same Confluent I got the following error
ERROR Failed to stream a record with null key!
(org.apache.ignite.stream.kafka.connect.IgniteSinkTask:98)
@SuppressWarnings("unchecked")
@Override public void put(Collection<SinkRecord> records) {
try {
for (SinkRecord record : records) {
if (record.key() != null) {
// Data is flushed asynchronously when
CACHE_PER_NODE_DATA_SIZE is reached.
StreamerContext.getStreamer().addData(record.key(),
record.value());
}
else {
log.error("Failed to stream a record with null key!");
}
}
}
catch (ConnectException e) {
log.error("Failed adding record", e);
throw new ConnectException(e);
}
}
After some analysis I found that this problem might be occurred since I have
not provided the properties like
--property parse.key=true --property key.separator=,
My question here is, Is it possible to pass this properties in kafka's
worker configuration file like connect.standalone ?
Thanks in advance !!
Austin
--
View this message in context:
http://apache-ignite-users.70518.x6.nabble.com/Kafka-Failed-to-stream-a-record-with-null-key-tp8731.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.
