Re: parallel processing - splitting data

If you have 4 Nodes with RF 4 then all data is on every node. So you can just slice the whole token range into 4 pieces and let each node process 1 slice. Determining local ranges also only helps if you read with CL_ONE. 2017-01-19 13:05 GMT+01:00 Frank Hughes : > Hello

Re: parallel processing - splitting data

I have tried to retrieve the token range and slice in 4, but the response i get for the following code is different on each node: TokenRange[] tokenRanges = unwrapTokenRanges(metadata.getTokenRanges(keyspaceName, localHost)).toArray(new TokenRange[0]); On each node, the 1024 token ranges are

Re: parallel processing - splitting data

I meant the global whole token range which is -(2^64/2) to ((2^64) / 2 - 1) I remember there are classes that already generate the right slices but don't know by heart which one it was. 2017-01-19 13:29 GMT+01:00 Frank Hughes : > I have tried to retrieve the token range

parallel processing - splitting data

Hello there, I'm running a 4 node cluster of Cassandra 3.9 with a replication factor of 4. I want to be able to run a java process on each node only selecting a 25% of the data on each node, so i can process all of the data in parallel on each node. What is the best way to do this with the java

Re: parallel processing - splitting data

Hi Frank, You could try this https://github.com/siddv29/cfs I have processed 1.2 billion rows in 480 seconds with just 20 threads on client side. C* 3.0.9 Nodes = 6 RF = 3 Have a go at it. You might be surprised. Regards, On Thu, Jan 19, 2017 at 5:35 PM, Frank Hughes

JVM state determined to be unstable. Exiting forcefully. what is Java Stability Inspector ?? why it is stopping DSE?

>From last few days i am seeing on some of the nodes in cassandra cluster DSE is getting shutdown due to the error below and i need to kill Java process and restart DSE service. I have cross checked reads and writes and compactions nothing looks suspicious, but i am seeing full Gc pause on these

Re: Unreliable JMX metrics

Thanks for you reply Malte. I replaced the nodes and the new nodes have new IP, but I have removed the dead nodes using nodetool. Does Cassandra still keep the info of dead nodes for 72 hours even I have removed them? Thanks, Guan From: Malte Pickhan

Re: JVM state determined to be unstable. Exiting forcefully. what is Java Stability Inspector ?? why it is stopping DSE?

Hi Pranay, what can be the reason for this It can be due to a JVM / GC misconfiguration or to some abnormal activity in Cassandra. Often, GC issues are a consequences and not the root cause of an issue in Cassandra. > how to debug that ?? how to fine grain why on those particular nodes this

Re: Unreliable JMX metrics

Yes. You likely will still be able to see the nodes in nodetool gossipinfo

3.0.8 AssertionError

The first time a client connects to the cluster and sends a bunch of inserts in parallel I get this assertion error. All subsequent inserts (including just retrying the first requests) work fine. The only assert in PrecisionTime is checking if the millis argument is <= the current millis. What

Re: JVM state determined to be unstable. Exiting forcefully. what is Java Stability Inspector ?? why it is stopping DSE?

what i have observed is 2-3 old gen GC's in 1-2 mins before OOM which i rarely see and seen hinted handoffs get accumulated on nodes which went down, and Mutation drops as well. i really don't know how to analyse hprof file is there any guide or blog that can help me how to analyse it ?? our