Database Reference
In-Depth Information
The default replication factor in a Kafka cluster is 1, meaning that only
the leader has the data. You can manually set the replication factor on a
per-topic basis by explicitly creating the topic, but this replication factor will
be used for all auto-created topics (if auto-creation is enabled).
# Set the default replication factor for a topic
#default.replication.factor=3
If a replica is being used, the next few parameters control the maintenance
of ISR sets. In general, there is no real reason to change most of these
defaults. (You should use care if you do change them.) In particular, setting
the lag time or messages to be too small can make it impossible for replicas
to remain in the ISR. The one setting that can sometimes improve
performance is increasing
num.replica.fetchers
to improve
replication throughput.
# The maximum time a leader will wait for a fetch
request before
# evicting a follower from the ISR
#replica.lag.time.max.ms=10000
# The maximum number of messages a replica can lag the
leader before it
# is evicted from the ISR
#replica.lab.max.messages=4000
# The socket timeout for replica requests to the leader
#replica.socket.timeout.ms=30000
# The replica socket receive buffer bytes
#replica.socket.receive.buffer.bytes=65536
# The maximum number of bytes to receive in each
replica fetch request
#replica.fetch.max.bytes=1048576
# The maximum amount of time to wait for a response
from the leader
# for a fetch request
#replica.fetch.wait.max.ms=500
# The minimum number of bytes to fetch from the leader
during a request
#replica.fetch.min.bytes=1
# The number of threads used to process replica
