java - kafka compression using high level consumer and simple consumer -
In my application, we are using a high level consumer who uses discommoded data without any problem, If manufacturers and consumers reject data using second and java API
the manufacturer uses the API (either SPP or GizIP) for librdkafka C ++ compression ? Can the Java consumer decontest transparently as it was done in the case of the above, what happens with the size of the fiber in the consumer side? Is it also handled transparently?
What happens if a consumer is made using a simple consumer model? Do we have to explicitly reject the compressed data coming from the manufacturer (assuming that librdkafka C ++ API is used here)
I'm thinking that compression In the case of Highlight, the client can not work with the librdkafka C ++ API in the productive side? Please explain to me if I am wrong here as I have seen some more posts here. Contradicting it, I have found another link that the decompression says to work when a high level consumer is used.
Thanks
They are compatible, compression similar to the Libridcafe Scala / Java client and Uses framing.
Enhance fetch.message.max.bytes Allows the user to bring large messages or leave large batches of messages with each request, but usually leave it at their default price You can be given as long as your producers do not make big messages from this value - in this case you will need to increase the message.max .bytes .
Compression is configured only on the manufacturer, since every message (or batch of message) has been flagged with the compression type (None, Fast, Zzip, ..) Configuration is not required.
Comments
Post a Comment