jaime
jaime
> One thing we could do is just to clean up all segment meta into new format when datacoord upgrade to new version, but it definitely take a very long...
Could you please refer this [script](https://github.com/milvus-io/milvus/tree/master/deployments/export-log) to export all Milvus logs?
This is log [client.log](https://github.com/confluentinc/confluent-kafka-go/files/8562925/client.log) that running on the dedicated cluster.
it will not occur hang after config message.max.bytes to 104857600, so the root cause seems to be caused when message.max.bytes on the client-side less than on the server-side.
@mhowlett it works well, I can get a "Message size too large" error when checking p.Produce return value
[log.txt](https://github.com/confluentinc/confluent-kafka-go/files/8676274/log.txt) there is debug log file
> Hi, @edenhill, it works after using this config, and I also try different configs multiple times, the root cause is the following configs: ``` "go.events.channel.size": 0, "go.produce.channel.size": 0, ```...