Re: Quorom consistency in a changing ring

2010-04-26 Thread David Timothy Strauss
Increasing the replication level is known to break it. --Original Message-- From: Peter Schuller Sender: sc...@scode.org To: user@cassandra.apache.org ReplyTo: user@cassandra.apache.org Subject: Quorom consistency in a changing ring Sent: Apr 26, 2010 21:55 Hello, Is my interpretation co

Re: Off line client nodes?

2010-04-12 Thread David Timothy Strauss
It's not common for me to recommend CouchDB, but this is one instance it's great for: synching complete datasets for disconnected use. Cassandra treats disconnection as a problem, not something that should occur in the normal plan of operations. -Original Message- From: Colin Yates Dat

Re: Integrity of batch_insert and also what about sharding?

2010-04-07 Thread David Timothy Strauss
Based on empirical usage, Gossip chatter is quite manageable well beyond 100 nodes. One advantage of many small nodes is that the cost of node failure is small on rebuild. If you have 100 nodes with a hundred gigs each, the price you pay for a node's complete failure is pulling a hundred gig

Re: Cassandra Design or another solution

2010-04-05 Thread David Timothy Strauss
Cache the => map as you write values (a "write-through" cache) so that reading the current score hits something like memcached instead of Cassandra. With a cache hit, you get an ideal, write-only path in Cassandra. Three blind writes in Cassandra is cheap -- no matter what your scale. The only

Re: Cassandra Design or another solution

2010-04-05 Thread David Timothy Strauss
te identification of the former 12th person to fill the 11th spot. (With a lack of monotonicity, the top eleven -- not just ten -- must always be tracked to efficiently know when someone drops below the top ten.) ----- "David Timothy Strauss" wrote: If user scores move in more than o

Re: Cassandra Design or another solution

2010-04-05 Thread David Timothy Strauss
If user scores move in more than one direction, as they apparently do in your case, they are not monotonic. Monotonicity can make system design a bit easier for various reasons. - "JKnight JKnight" wrote: Thanks David, But what's does "monotonicity" mean? User's score belongs to thei

Re: Best PHP client

2010-04-02 Thread David Timothy Strauss
I've written one here as a Drupal module, but it isn't very Drupal-specific: http://bazaar.launchpad.net/~davidstrauss/pressflow/cassandra-votingapi/files/head:/sites/all/modules/cassandra/ I think my interface is a little prettier than SimpleCassie. You can see it in action in the test suite

Re: Cassandra data file corrupt

2010-03-31 Thread David Timothy Strauss
Cassandra has always supported two great ways to prevent data loss: * Replication * Backups I doubt Cassandra will ever focus extensively on single-node recovery when it's so easy to wipe and rebuild any node from the cluster. -Original Message- From: JKnight JKnight Date: Wed, 31 Mar

Re: Replicating data over the wan?

2010-03-30 Thread David Timothy Strauss
Your ConsistencyLevel will change the effect. If CL is low, inconsistency will temporarily occur between the DCs. If CL is high, writes will have noticeably high latency. -Original Message- From: Erik Holstad Date: Tue, 30 Mar 2010 17:49:17 To: Subject: Replicating data over the wan?

Re: Poor performance; PHP & Thrift to blame

2010-03-30 Thread David Timothy Strauss
larger query sets. I suppose Facebook and Digg are only pulling out small column sets, so they wouldn't necessarily notice this issue. On Tue, Mar 30, 2010 at 8:00 PM, David Timothy Strauss wrote: > Without APC, there should be even more of an improvement with the Thrift PHP > exte

Re: Poor performance; PHP & Thrift to blame

2010-03-30 Thread David Timothy Strauss
Without APC, there should be even more of an improvement with the Thrift PHP extension. - "Rauan Maemirov" wrote: > What about APC? Did you turn it on? > > 2010/3/30 Julian Simon : > > Hi, > > > > I've been trying to benchmark Cassandra for our use case and have > been > > seeing poor perf

Re: Write times

2010-03-29 Thread David Timothy Strauss
OPP should only affect write speed if OPP's tendency to unevenly distribute load causes some nodes to be overworked. In other words, OPP vs. RP on a single node system should have no real effect. -Original Message- From: Carlos Sanchez Date: Mon, 29 Mar 2010 18:58:50 To: user@cassandra