2019-06-20 09:41:12 UTC - Allen ONeill: Hi all, question about functions..... Do they operate standalone as consumers (the diagrams seem to indicate this), or are they inline processes that can peek/update messages before they are notified to standard consumers? ---- 2019-06-20 13:13:56 UTC - Fredrick P Eisele: re: pulsar-client : I am working with the standalone cluster. When using ./bin/pulsar-client I noticed the old style topic containing the cluster e.g. <persistent://sample/standalone/ns1/hello-world> rather than <persistent://sample/ns1/hello-world>. And ./bin/pulsar-admin namespaces list sample sample/standalone/ns1 Is this a backward compatibility thing? It seems namespaces can contain slashes? ---- 2019-06-20 13:19:48 UTC - Shivji Kumar Jha: We have productionized pulsar just this week(2.3.1 version). The client connections work well but I see this crash in my log time and again. Am I missing an auth config or something? ---- 2019-06-20 13:20:38 UTC - Shivji Kumar Jha: ---- 2019-06-20 13:31:24 UTC - Shivji Kumar Jha: @Matteo Merli maybe this is related to <https://github.com/apache/pulsar/pull/3869/files> ? Just guessing because both have redirect context… ---- 2019-06-20 13:36:30 UTC - Shivji Kumar Jha: Another symptom, maybe same cause? ---- 2019-06-20 13:37:10 UTC - Shivji Kumar Jha: ---- 2019-06-20 13:52:08 UTC - Brian Doran: @Sijie Guo/@Penghui Li Any update on the 2.4.0 release? ---- 2019-06-20 15:27:55 UTC - Penghui Li: i will start tomorrow ---- 2019-06-20 15:29:07 UTC - Aaron: @Sijie Guo ---- 2019-06-20 15:29:19 UTC - Aaron: @Sijie Guo ---- 2019-06-20 15:50:43 UTC - Prakash Saraf: Hi, I am having a question w.r.t schema in pulsar. If the topic has messages produced by using two different version of schema, how can a consumer consume both old as well as new mesgs? ---- 2019-06-20 16:46:29 UTC - Shivji Kumar Jha: Whats your schema format (avro, json, protobuff, bytes ) ? ---- 2019-06-20 17:44:36 UTC - Prakash Saraf: avro ---- 2019-06-20 17:57:50 UTC - Prakash Saraf: seems its fixed in 2.4 - <https://github.com/apache/pulsar/pull/3876> ---- 2019-06-20 18:21:05 UTC - Matteo Merli: @Shivji Kumar Jha This is related to broker->broker authentication.
When a bundle (eg: a shard of a namespace) meets a certain threshold of traffic, it will be candidate to get split, so that the 2 subsets of topics can be moved to different brokers. The error is that this split fails because the broker is not passing the Auth credentials when talking to other brokers. Take a look at : ``` brokerClientAuthenticationPlugin= brokerClientAuthenticationParameters= ``` This should be set such that brokers have “super-user” credentials clap : Shivji Kumar Jha ---- 2019-06-20 18:26:45 UTC - Guillaume Rosauro: @Guillaume Rosauro has joined the channel ---- 2019-06-20 18:32:44 UTC - Guillaume Rosauro: I have a question related to debezium CDC connector. I do not really understand how it works. I mean ```bin/pulsar-admin source localrun --sourceConfigFile debezium-postgres-source-config.yaml``` seems to start debezium/kafka/zookeeper, but I am not really sure about that. So how pulsar <-> debezium are linked and work together ? An other question related to kafka, with Debezium, if I want to configure to configure table_name => kafka topic mapping, I would configure kafka like this: ``` "transforms":"dropPrefix", "transforms.dropPrefix.type":"org.apache.kafka.connect.transforms.RegexRouter", "transforms.dropPrefix.regex":"DC1-TEST-(.*)", "transforms.dropPrefix.replacement":"$1" ``` But how could this be done with pulsar ? ---- 2019-06-20 18:56:46 UTC - Ali Ahmed: @Guillaume Rosauro You can reference the settings here <https://debezium.io/blog/2019/05/23/tutorial-using-debezium-connectors-with-apache-pulsar/> ---- 2019-06-20 19:07:10 UTC - Guillaume Rosauro: Thanks. I have already used this tutorial and it is working fine (for 2.3; for 2.4 you need some small modifications). But I still can not understand why it is working as I did not start Debezium first, but Pulsar seems to start it by itself. ---- 2019-06-20 21:36:21 UTC - Sijie Guo: Each message is associated with a schema version. Currently the struct schemas (e.g. avro, json) are taking care of that already. You don’t need to worry about that. If you want to get the schema info for a specific version, you can use the admin schemas api to retrieve the schema info for a given schema version ---- 2019-06-20 21:36:51 UTC - Sijie Guo: I believe the unacked messages will be redelivered twice. ---- 2019-06-20 21:58:47 UTC - Prakash Saraf: @Sijie Guo - with changes done as part of <https://github.com/apache/pulsar/pull/3876> will the consumer be able to read messages generated with different version of AVRO schemas? ---- 2019-06-20 22:03:14 UTC - Sijie Guo: @Prakash Saraf I actually just wrote the reply to your email :slightly_smiling_face: ---- 2019-06-20 22:03:35 UTC - Sijie Guo: and yes. that is the change you need for the problem you hit. ---- 2019-06-20 22:06:14 UTC - Prakash Saraf: thank you!! looking forward to release.. I will try to compile from master tomorrow or early next week. ---- 2019-06-21 02:59:02 UTC - Mahesh: @Mahesh has joined the channel ----
