Closed Dwijad closed 5 years ago
hi @DwijadasDey It's a bit tough to exactly pin point the problem you are facing given the above information...
1) as for the plugin-path this is typically the "main entry" folder where you want the classpath scanning for connector implementations to happen
2) the idea is also to keep every connector inside a separate sub-folder in order to allow for classpath isolation during bootstrap face of connect. here it looks like you are mixing all jars into one folder which belong to different connectors?
The following are comments in a sample connect-distributed.properties
file which for instance is taken from the Confluent Platform:
# Set to a list of filesystem paths separated by commas (,) to enable class loading isolation for plugins
# (connectors, converters, transformations). The list should consist of top level directories that include
# any combination of:
# a) directories immediately containing jars with plugins and their dependencies
# b) uber-jars with plugins and their dependencies
# c) directories immediately containing the package directory structure of classes of plugins and their dependencies
the errors you see may thus result form other connector jars that are also in the plugin-path. the verify this you could remove all the other connectors and only put the mongodb sink connector into your plugin-path and restart connect. you may want to try if if helps to move the complete folder "hpgrahsl-kafka-connect-mongodb-1.3.1" that you get after unzipping into the plugin-path entry folder.
irrespective of this missing dependencies, the error message you get in terms of a 500 "no route to host" is definitely something else and you need to dig deeper and also get more logs which refer to this error.
Hi @hpgrahsl I had to configure the connector in the rest of the nodes in the cluster to remove the error messages. Earlier i tried to run the connector by configuring the connector in a single node. Although, I am not sure whether this is the reason for the offending error messages or not.
Earlier even though there were error messages in the startup but the node was able to initialized with other connectors.
Regards
@DwijadasDey THX for reporting back! So from your answer I conclude that you solved your problem which was that you only installed the sink connector on one of several nodes of your distributed connect deployment. If I'm right I'll close this issue. If not feel free to reopen anytime.
Hi I am facing an issue while trying to start a worker/connector(sink). Attaching the startup log as an attachment. I have configured the connector like following:
plugin.path=/u01/cnfkc/confluent-kfk/share/java/kafka-connect-mongodb
When i try to start the
connect-distributed
using the above properties file, there are lot of class not found errors starting with the following one:The plugin can also be listed through REST call
Submitting a connector through REST results in following error which i think cannot connect to MongoDB host although i can ping and connect to the mongodb host/port.
STartup log
It seems all exceptions are related to Reflections.
I have the following JARS in the plugin path
Shall i have to include more JAR file in the plugin path ?