librdkafka

librdkafka copied to clipboard

The sequence diagram for the consumer group appears to have an error on the second rebalance the coordinator seems more likely to return a response than a request. 

fgmarand

create new topic-partition list comparison method based on hashing

2

the new method is much faster than the old, O(Na + Nb) compared to the older one, which is O(Na * Nb) Background: The old one was causing problems with...

lucas-sonnabend

[DOCS]: Update Node.js language binding link

1

This update corrects and point the link to the confluent supported wrapper library for node.js.

jaypeeig

Disconnected while requesting ApiVersion: might be caused by incorrect security.protocol configuration (connecting to a SSL listener?) or broker version is < 0.10

1

Hello, I've been trying to connect to the Azure Event Hub using the lib and getting the error constantly. I've tried different options and still no go. Is there any...

kefiras

rd_kafka_new call returns after 2 minutes and 30 seconds

1

While investigating an issue where our Kafka producer failed to initialize on a specific machine, we discovered something interesting. Although the root problem is likely with the machine itself, I...

whyjp

Memory Leak Issue Found When Using rdkafka Library

1

Hello! During the process of using the rdkafka library, I detected a memory leak issue through the Valgrind tool. Here are the detailed information: ``` ==2495== 16 bytes in 1...

Huu-Yuu

The Kafka is deployed in a kubernetes environment with 3 pods in a kafka cluster. The producer kafka connection keeps flapping. The error callback is producing below low - **err:...

mukulksrivastava

[KIP-848] Fix static group membership

1

there are three cases for static group membership this PR fixes the first one: - member explicitly unsubscribes -> should cause a rebalance - member closes the consumer -> no...

emasab

Since: first_seq has int32_t type https://github.com/confluentinc/librdkafka/blob/93877617709eb071a0f4ec7038c54e2764abefc9/src/rdkafka_msgbatch.h#L37 and rd_kafka_msgq_len return int(int32) https://github.com/confluentinc/librdkafka/blob/93877617709eb071a0f4ec7038c54e2764abefc9/src/rdkafka_msg.h#L244-L246 But rd_kafka_seq_wrap accepts int64_t as argument https://github.com/confluentinc/librdkafka/blob/93877617709eb071a0f4ec7038c54e2764abefc9/src/rdkafka_msg.h#L594-L596 There could be overflow at https://github.com/confluentinc/librdkafka/blob/93877617709eb071a0f4ec7038c54e2764abefc9/src/rdkafka_request.c#L4655-L4656 If there's no overflow then what's...

stasos24

Since inflight_cnt, retry_cnt, outq_cnt has int type https://github.com/confluentinc/librdkafka/blob/93877617709eb071a0f4ec7038c54e2764abefc9/src/rdkafka_broker.c#L941-L942 And rd_atomic64_add accepts two int64 https://github.com/confluentinc/librdkafka/blob/93877617709eb071a0f4ec7038c54e2764abefc9/src/rdatomic.h#L144-L145 There is possible integer overflow at: https://github.com/confluentinc/librdkafka/blob/93877617709eb071a0f4ec7038c54e2764abefc9/src/rdkafka_broker.c#L969-L970 Possible fix: `(int64) inflight_cnt + outq_cnt`

stasos24