Hello, I'm playing a bit with Cassandra and had a look at the code to understand better how reads and writes works.
And I have a question about the weak read protocol . If I understand correctly, a weak read either read the row locally (if that makes sense) or pick the "closest" node and ask him to send its row. In the case of a remote weak read, if the node timeout, the whole read timeout. Wouldn't that be worth it, for availability sake, to ask another replica ? Well, if I say that, it's because it is the behavior the comments of the StorageProxy.weakReadRemote method actually suggests. But it doesn't seems to be what the code does. Am I completely mistaking ? If so, sorry about it. And while I'm at it, the comments of the function weakReadLocal says that if the local row is NULL, it'll do the read from another replica. But again, that doesn't seems to be the case. Again, I'm fairly new to Cassandra and its code, so if I'm wrong, sorry. But otherwise, the discrepancy between code and comments suggests that the code was changed. If so, what was the rational behind the change ? Thanks, Sylvain
