[ https://issues.apache.org/jira/browse/KAFKA-5938?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16176716#comment-16176716 ]
Robin Moffatt commented on KAFKA-5938: -------------------------------------- Just to confirm, this does work against Oracle with Connect in distributed mode. I couldn't get your Docker example to work (I'm not a Docker expert; the networking didn't seem to work for me, as none of the cited addresses worked). I ran this against my machine running Oracle 12.2 and Confluent Platform 3.3 {code} curl -X "POST" "http://localhost:8082/connectors/" -H "Content-Type: application/json" -d $'{ "name": "countries", "config": { "connection.url": "jdbc:oracle:thin:@localhost:1521/orcl", "connection.password": "oracle", "topic.prefix": "countries", "mode": "bulk", "connector.class": "io.confluent.connect.jdbc.JdbcSourceConnector", "tasks.max": 1, "query": "select * from HR.COUNTRIES", "connection.user": "SYSTEM", "poll.interval.ms": 100000 } }' {code} The code is the same as the one you quote, except I've modified the JDBC URL for my host and SID. The connector is created and runs successfully: {code} $ confluent status connectors| jq '.[]'| xargs -I{connector} confluent status {connector}| jq -c -M '[.name,.connector.state,.tasks[].state]|join(":|:")'| column -s : -t| sed 's/\"//g'| sort countries | RUNNING | RUNNING {code} The topic is created and populated: {code} $ kafka-topics --zookeeper localhost:2181 --list [...] countries [...] $ kafka-avro-console-consumer \ --bootstrap-server localhost:9092 \ --property schema.registry.url=http://localhost:8081 \ --from-beginning \ --topic countries \ --max-messages 1 [2017-09-22 12:51:15,012] WARN Ignoring invalid logical type for name: decimal (org.apache.avro.LogicalTypes:95) {"COUNTRY_ID":"AR","COUNTRY_NAME":{"string":"Argentina"},"REGION_ID":{"bytes":"v.Ôg!=ÑòÅZg_4÷) °úoÏÑ\u000BeÍ\u001E·rÍ\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000"}} Processed a total of 1 messages oracle@localhost ~> {code} You'll note that there's a datatype warning and the message contents itself appears scrambled -- but that is a separate issue from the one reported here. One other thought, this is on a single-node Distributed instance; have you tried that already? Do you have the Oracle JDBC driver present on each of the Connect workers that you're running in the distributed cluster? As an observation, using the configuration option {{query}} for a {{SELECT *}} seems somewhat superfluous when you could just use {{table.whitelist}} instead and not need to specify the query at all. > Oracle jdbc-source-connector with kafka-connect distributed mode will result > in http 500 error > ---------------------------------------------------------------------------------------------- > > Key: KAFKA-5938 > URL: https://issues.apache.org/jira/browse/KAFKA-5938 > Project: Kafka > Issue Type: Bug > Components: KafkaConnect > Affects Versions: 0.11.0.0 > Reporter: Wolfgang Clemens > > When I try to add an oracle jdbc source connector in kafka-connect > distributed mode via curl it will end up in a http 500 error. > To reproduce, feel free to clone my repo for that error: > https://github.com/TranceMaster86/kafka-connect-oracle-error > In the readme I described how to reproduce the error. > It will hang up kafka-connect completely. All other connectors seems to stop > working after I want to create an oracle jdbc source connector with a '@' in > the connection string. I think this will be the main problem here. The oracle > connection string looks like 'jdbc:oracle:thin:@hostname:port/service' or > 'jdbc:oracle:thin:@hostname:port:SID'. > See also: https://github.com/confluentinc/cp-docker-images/issues/338 -- This message was sent by Atlassian JIRA (v6.4.14#64029)