[CASSANDRA-5632] Cross-DC bandwidth-saving broken

• Type: Bug
• Status: Resolved
• Priority: Normal
• Resolution: Fixed
• Fix Version/s: 1.2.6
• Severity: Normal
• Assignee: Jonathan Ellis
• Reporter: Jonathan Ellis

Authors

Jonathan Ellis

Reviewers

David Brosius

• Tester: Hayato Shimizu
• Created: 12/Jun/13 23:37
• Updated: 16/Apr/19 09:32
• Resolved: 20/Jun/13 20:54

Description

We group messages by destination as follows to avoid sending multiple messages to a remote datacenter:

        // Multimap that holds onto all the messages and addresses meant for a specific datacenter
        Map<String, Multimap<Message, InetAddress>> dcMessages

When we cleaned out the MessageProducer stuff for 2.0, this code

                    Multimap<Message, InetAddress> messages = dcMessages.get(dc);
...
                    messages.put(producer.getMessage(Gossiper.instance.getVersion(destination)), destination);

turned into

                    Multimap<MessageOut, InetAddress> messages = dcMessages.get(dc);
...
                    messages.put(rm.createMessage(), destination);

Thus, we weren't actually grouping anything anymore – each destination replica was stored under a separate Message key, unlike under the old CachingMessageProducer.

Issue Links

• is duplicated by: CASSANDRA-6314 Cassandra sends multiple copies of data over WAN