[ https://issues.apache.org/jira/browse/CASSANDRA-8193?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14207377#comment-14207377 ]
Yuki Morishita commented on CASSANDRA-8193: ------------------------------------------- First of all, thanks for the patch! I review it based on 2.0, but because the patch adds new feature, I'd rather put this to 2.1+. (So go ahead and apply 2.0.x yourself after review). So, some comments: * If replication factor is set to be 1 for each DC, then it will be the same as ParallelRequestCoordinator. There needs fall back to current behavior in this case. * It looks like ParallelRequestCoordinator class can be {{... implements IRequestCoordinator<InetAddress>}}. * DatacenterAwareRequestCoordinator uses AtomicInteger, but primitive int just works here. * nit: put braces on a new line. > Multi-DC parallel snapshot repair > --------------------------------- > > Key: CASSANDRA-8193 > URL: https://issues.apache.org/jira/browse/CASSANDRA-8193 > Project: Cassandra > Issue Type: Improvement > Reporter: Jimmy Mårdell > Assignee: Jimmy Mårdell > Priority: Minor > Fix For: 2.0.12 > > Attachments: cassandra-2.0-8193-1.txt > > > The current behaviour of snapshot repair is to let one node at a time > calculate a merkle tree. This is to ensure only one node at a time is doing > the expensive calculation. The drawback is that it takes even longer time to > do the merkle tree calculation. > In a multi-DC setup, I think it would make more sense to have one node in > each DC calculate the merkle tree at the same time. This would yield a > significant improvement when you have many data centers. > I'm not sure how relevant this is in 2.1, but I don't see us upgrading to 2.1 > any time soon. Unless there is an obvious drawback that I'm missing, I'd like > to implement this in the 2.0 branch. -- This message was sent by Atlassian JIRA (v6.3.4#6332)