Looks like a bytebuf issue. When reading a bytebuf, we often do things like reading the size from the packet and setting the writeIndex to the current writeIndex + size, so that it can be passed into the protobuf parser. This is why the write index seems so messed up. Do you have a reliable repro, or do you just see it happening regularly in production?
Is TLS auth enabled, or just TLS transport? this is going to be a bytebuf pooling issue of some sort with TLs. [ Full content available at: https://github.com/apache/incubator-pulsar/issues/2401 ] This message was relayed via gitbox.apache.org for [email protected]
