Re: Using a synchronized counter that keeps track of no of users on the application & using it to allot UserIds/ keys to the new users after sign up

2011-02-28 Thread Ertio Lew
On Tue, Mar 1, 2011 at 1:26 AM, Aaron Morton wrote: > This is mostly from memory. But the last 12 ? (4096 decimal) bits are a > counter for the number of id's generated in a particular millisecond for that > server. You could use the high 4 bits in that range for your data type flags > and the

Re: Question about insert performance in multiple node cluster

2011-02-28 Thread Ryan King
On Mon, Feb 28, 2011 at 2:05 PM, Flachbart, Dirk (HP Software - TransactionVision) wrote: > Replication factor is set to 1, and I'm using ConsistencyLevel.ANY. And yep, > I tried doubling the threads from 16 to 32 when running with the second > server, didn't make a difference. > > Regarding the

Re: Question about insert performance in multiple node cluster

2011-02-28 Thread Peter Schuller
> Replication factor is set to 1, and I'm using ConsistencyLevel.ANY. And yep, > I tried doubling the threads from 16 to 32 when running with the second > server, didn't make a difference. Are you sure the client isn't the bottleneck? Have you tried running the client on independent (and perhaps

RE: Question about insert performance in multiple node cluster

2011-02-28 Thread Flachbart, Dirk (HP Software - TransactionVision)
Nope, I'm on a Gigabit network. The windows task manager on both machines shows a network utilization of around 12 percent. Regards, Dirk -Original Message- From: sc...@scode.org [mailto:sc...@scode.org] On Behalf Of Peter Schuller Sent: Monday, February 28, 2011 12:53 PM To: user@cassa

RE: Question about insert performance in multiple node cluster

2011-02-28 Thread Flachbart, Dirk (HP Software - TransactionVision)
Replication factor is set to 1, and I'm using ConsistencyLevel.ANY. And yep, I tried doubling the threads from 16 to 32 when running with the second server, didn't make a difference. Regarding the ring balancing - I assume it should be balanced. I'm using RandomPartitioner, and the keys are gen

Time to rebuild a node

2011-02-28 Thread A J
Hello, I know it depends on lot of factors but are there any ballpark number on how long it takes to recreate a node from other nodes (or add a new node to a ring). Something like x GBs take y minutes to build ? Thanks.

Re: Using a synchronized counter that keeps track of no of users on the application & using it to allot UserIds/ keys to the new users after sign up

2011-02-28 Thread Aaron Morton
This is mostly from memory. But the last 12 ? (4096 decimal) bits are a counter for the number of id's generated in a particular millisecond for that server. You could use the high 4 bits in that range for your data type flags and the low 8 for the counter. Aaron On 1/03/2011, at 4:41 AM, Ert

Re: node failure, and automatic decommission (or removetoken)

2011-02-28 Thread Aaron Morton
I thought there was more to it. The steps for move or removing nodes are outlined on the operations page wiki as you probably know. What approach are you considering to rebalancing the token distribution when removing a node? E.g. If you have 5 nodes and remove 1 the best long term solution is

Re: Column family cannot be removed

2011-02-28 Thread Jonathan Ellis
drop and truncate both snapshot first, which requires forking to run ln if you don't have JNA installed. best solution: install JNA so it can do in-process link calls. On Mon, Feb 28, 2011 at 6:05 AM, George Ciubotaru wrote: > Hello guys, > > > > I’m trying to remove a column family but without

Re: Question about insert performance in multiple node cluster

2011-02-28 Thread Peter Schuller
> What's your replication factor? Which consistency level are you using? > Is the ring evenly balanced? Did you double the number of client > threads when you added the second server? And are you on 100 mbit networking? 9k requests/second inserting 1k, sounds suspiciously close to saturating 100 M

Re: ColumnFamilyRecordWriter

2011-02-28 Thread Jeremy Hanna
One thing that could be done is the CFRW could be abstracted more so that it's easier to extend and only the serialization mechanism is required to extend it. That is, all of the core functionality relating to Cassandra would be in an abstract class or something like that. Then the avro based

Re: Question about insert performance in multiple node cluster

2011-02-28 Thread Ryan King
On Mon, Feb 28, 2011 at 9:24 AM, Flachbart, Dirk (HP Software - TransactionVision) wrote: > Hi, > > > > We are trying to use Cassandra for high-performance insertion of simple > key/value records. I have set up Cassandra on two of my machines in my local > network (Windows 2008 server), using pret

Question about insert performance in multiple node cluster

2011-02-28 Thread Flachbart, Dirk (HP Software - TransactionVision)
Hi, We are trying to use Cassandra for high-performance insertion of simple key/value records. I have set up Cassandra on two of my machines in my local network (Windows 2008 server), using pretty much the default configuration. I created a test driver in java (using thrift) which inserts a sin

Re: ColumnFamilyRecordWriter

2011-02-28 Thread Jeremy Hanna
There certainly could be a thrift based record writer. However, (if I remember correctly) to enable Hadoop output streaming, it was easier to go with Avro for doing the records as the schema is included. There could also have been a thrift version of the record writer, but it's simpler to just

Re: Using a synchronized counter that keeps track of no of users on the application & using it to allot UserIds/ keys to the new users after sign up

2011-02-28 Thread Ertio Lew
Hi Ryan, I am considering snowflake as an option for my usage with Cassandra for a distributed application. As I came to know snowflake uses 64 bits IDs. I am looking for a solution that could help me generate 64 bits Ids but in those 64 bits I would like at least 4 free bits so that I could manip

Re: node failure, and automatic decommission (or removetoken)

2011-02-28 Thread Mimi Aluminium
Aaron, Thanks a lot, Actually I meant a larger number of nodes than 3 and replication factor of 3. We are looking on a system that may shrink due to permanent failures, and then automatically detects the failure and stream its range to other nodes in the cluster to have again 3 replicas. I understn

Column family cannot be removed

2011-02-28 Thread George Ciubotaru
Hello guys, I'm trying to remove a column family but without success. I have a simple 3 nodes Cassandra 0.7 cluster, a keyspaces with replication factor 2 and a super column within this keyspace with around 200,000 rows. I'm using cassandra-cli: - drop column family CFName; throws the

Re: How to restrict access to cassandra jmx to private network?

2011-02-28 Thread ruslan usifov
2011/2/28 ruslan usifov > > 2). Add folow lines into configuration: > > -Dcom.playrix.cassandra.management.port=8080 > -Dcom.playrix.cassandra.management.host= > -Djava.rmi.server.hostname= > -javaagent:"/usr/share/java/cassandra-agent.jar" > > This is wrong settings, here is right: -Dorg.apach

Re: How to restrict access to cassandra jmx to private network?

2011-02-28 Thread ruslan usifov
Thanks Matt, as result of my investigations is simple javaganet that, bind to specific network interface, and I want share it with community (if it will be interested). How to use: 1). You must disable standard jmx. (i.e. remove all com.sun.management.jmxremote.* settings) 2). Add folow lines int

ColumnFamilyRecordWriter

2011-02-28 Thread Mayank Mishra
Hi all, As I was integrating Hadoop with Cassandra, I wanted to serialize mutations, hence I used thrift mutations in M/R jobs. During the course, I came to know that CFRW considers only Avro mutations. Can someone please explain me why only avro transport is entertained by CFRW. Why not, bo

Re: node failure, and automatic decommission (or removetoken)

2011-02-28 Thread aaron morton
AFAIK the general assumption is that you will want to repair the node manually, within the GCGraceSeconds period. If this cannot be done then nodetool decomission and removetoken are the recommended approach. In your example though, with 3 nodes and an RF of 3 your cluster can sustain a single

Re: How to restrict access to cassandra jmx to private network?

2011-02-28 Thread aaron morton
Take a look at the bottom of conf/cassandra-env.sh and follow the link to http://blogs.sun.com/jmxetc/entry/troubleshooting_connection_problems_in_jconsole I think you can bind it to an interface using -Djava.rmi.server.hostname JVM option Aaron On 28 Feb 2011, at 04:22, Matt Kennedy wrote: