collector cores in kafka seen r4.0.1.0-68 newton

[wenqing liang]

(gdb) bt
#0 0x00007f3507335b96 in strlen () from /lib/x86_64-linux-gnu/libc.so.6
#1 0x0000000000533456 in rd_kafka_assignor_cmp_str () at /usr/include/boost/exception/exception.hpp:78
#2 0x0000000000536848 in rd_list_find () at /usr/include/boost/exception/exception.hpp:78
#3 0x0000000000533ae6 in rd_kafka_assignor_run () at /usr/include/boost/exception/exception.hpp:78
#4 0x000000000051bde4 in rd_kafka_cgrp_assignor_run () at /usr/include/boost/exception/exception.hpp:78
#5 0x000000000051bf2c in rd_kafka_cgrp_assignor_handle_Metadata_op () at /usr/include/boost/exception/exception.hpp:78
#6 0x0000000000509aac in rd_kafka_op_call () at /usr/include/boost/exception/exception.hpp:78
#7 0x000000000050a8f5 in rd_kafka_op_handle () at /usr/include/boost/exception/exception.hpp:78
#8 0x000000000050588d in rd_kafka_q_serve () at /usr/include/boost/exception/exception.hpp:78
#9 0x00000000004e162c in rd_kafka_thread_main () at /usr/include/boost/exception/exception.hpp:78
#10 0x0000000000536a77 in _thrd_wrapper_function () at /usr/include/boost/exception/exception.hpp:78
#11 0x00007f3508b3b6ba in start_thread () from /lib/x86_64-linux-gnu/libpthread.so.0
#12 0x00007f35073b182d in clone () from /lib/x86_64-linux-gnu/libc.so.6
(gdb)

root@CB-R4-0-u16-newton-10-84-35-242:~# dpkg -l | grep librdka
ii librdkafka++1:amd64 0.11.0-0contrail0 amd64 library implementing the Apache Kafka protocol (C++ bindings)
ii librdkafka-dev:amd64 0.11.0-0contrail0 amd64 library implementing the Apache Kafka protocol (development headers)
ii librdkafka1:amd64 0.11.0-0contrail0 amd64 library implementing the Apache Kafka protocol

The cores are uploaded to /cs-shared/bugs/

[Anish Mehta (amehta00)]

Please confirm if this is reproducible (build 69)

[Arvind (arvindv)]

Seems like a driver issue. When partitions are added to consumer, the library is crashing
https://github.com/confluentinc/confluent-kafka-go/issues/71

2017-09-25 Mon 17:24:00:226.719 UTC server6 [Thread 139865552930560, Pid 11888]: LOG-5-PARTCNT: [thrd:main]: Topic -uve-agg-UveVirtualNetworkAgent-ingress_flow_count partition count changed from 20 to 30

Message seen before the crash on the collector.

[Anish Mehta (amehta00)]

We do not plan to address until next kafka upgrade.
Please let us know if this is seen often.