Panic when a message gets deleted (using a retention period) while consuming about kt HOT 2 CLOSED

hi @utdemir - sorry, missed this one. i just tried to reproduce it with the following steps:

$ git checkout v11.1.0
$ go install .
$ ./kafka-topics.sh --create --topic zzz --partitions 1 --replication-factor 1 --zookeeper zookeeper:2181

then keep producing:

while true; do echo '{"key": "1", "value": "hans"}' | kt produce -topic zzz ; sleep 1 ; done

and consuming:

kt consume -topic zzz

and then change the config:

$ ./kafka-configs.sh --zookeeper zookeeper:2181 --alter --entity-type topics --add-config 'retention.ms=10, cleanup.policy=delete' --entity-name "zzz"

but i don't see the nil pointer dereference - am i missing a step?

reading the relevant lines - v11.1.0 tried to access the contents of a message even if the chan was closed, so i'm guessing that's what you encountered. master doesn't have that ordering issue and i added a case to listen on the partition consumer's err chan. if you can reliably reproduce the issue, could you try master?

from kt.

hi @utdemir - i'll close this optimistically for now - please feel free to re-open if you can reproduce it with master :)

from kt.