For question 2, you can't inform the coordinator that you want it to use a certain node to fulfill the answer. It'll ask several and use the first count which answer that satisfy the consistency constraints.
On Fri, Sep 2, 2016, 1:19 PM Siddharth Verma <verma.siddha...@snapdeal.com> wrote: > I am debugging an issue at our cluster. Trying to find the RCA of it, > according to our application behavior. > Used WhiteList policy(asked a question for the same, some time back) but, > it was stated that it can not guarantee the desired behavior. > Yes, I forgot to mention, i was referring to Java driver. > I used DCAwareRoundRobin, TokenAware Policy for application flow. > Would ask question1 on driver mailing list, If someone could help with > question 2. > > > > > > > > > > > On Fri, Sep 2, 2016 at 6:59 PM, Eric Stevens <migh...@gmail.com> wrote: > >> These sound like driver-side questions that might be better addressed to >> your specific driver's mailing list. But from the terminology I'd guess >> you're using a DataStax driver, possibly the Java one. >> >> If so, you can look at WhiteListPolicy if you want to target specific >> node(s). However aside from testing specific scenarios (like performance >> testing coordinated operations) it's unlikely that with a correctly tuned >> LBP, you'll be able to do a better job of node selection than the driver is >> able to. DCAwareRoundRobin with a child policy of TokenAware will choose a >> primary or replica node for any operations where it can know in advance. >> >> With RF == N like your setup, every piece of data is owned by every node, >> so as long as your LBP is distributive, and outside of performance testing, >> I can't see why you'd be needing to target specific nodes for anything. >> >> On Fri, Sep 2, 2016 at 1:59 AM Siddharth Verma < >> verma.siddha...@snapdeal.com> wrote: >> >>> Hi, >>> I have Dc1(3 nodes), Dc2(3 nodes), >>> RF:3 on both Dcs >>> >>> question 1 : when I create my LoadBalancingPolicy, and override >>> newQueryPlan, the list of hosts from newQueryPlan is the candidate >>> coordinator list? >>> >>> question 2 : Can i force the co-ordintor to hit a particular cassandra >>> node only. I used consistency LOCAL_ONE, but i guess, i doesn't guarantee >>> that data will be fetched from it. >>> >>> Thanks >>> Siddharth Verma >>> >> >
