RE: Extremly slow inserts on LAN

We had similar experience. Problem was with TSocket as transport alone: var transport = new TSocket("192.168.0.123", 9160); var protocol = new TBinaryProtocol(transport); var client = new Cassandra.Client(protocol); Using TBufferedTransport helped a lot: var tran

RE: Extremly slow inserts on LAN

No - just Windows. So I'm going to do some experiments to isolate the cause: - use java client on windows - use linux server - use java client on linux Thanx -Original Message- From: David Strauss [mailto:da...@fourkitchens.com] Sent: May 09, 2010 5:48 PM To: user@cassandra.apache.org Su

Re: Cassandra training on May 21 in Palo Alto

In my opinion, To share the training material for us who do not living in the USA is a good idea. Thanks. Dir On Mon, May 10, 2010 at 2:25 AM, S Ahmed wrote: > I guess the hard part would be recording something so long (9-5pm) > > A video that is split between the screen (say powerpoint) and li

Re: Cassandra training on May 21 in Palo Alto

I guess the hard part would be recording something so long (9-5pm) A video that is split between the screen (say powerpoint) and linux console would be perfect :) On Fri, May 7, 2010 at 11:24 AM, Todd Burruss wrote: > +1 > > > > -Original Message- > *From:* S Ahmed [sahmed1...@gmail.co

Re: key is sorted?

True, The Range slice support was enabled in Random Partitioner for the hadoop support. Random partitioner actually hash the Key and those keys are sorted so we cannot have the actual key in order (Hope this doesnt confuse you)... Regards, On Sun, May 9, 2010 at 12:00 AM, David Boxenhorn

Re: Is SuperColumn necessary?

I'm not sure this is much of an improvement. It does illustrate, however, the desire to couch the concepts in terms that each is already comfortable with. Nearly every set of terms which come from an existing system will have baggage which doesn't map appropriately. Not that the "sparse multidimens

Re: key is sorted?

just to correct myself: to sort timeUUID on column in descending order, sliceRange.setReversed(true) does the trick. -aj On Sun, May 9, 2010 at 9:07 AM, AJ Chen wrote: > true, i think. > but don't rush to OrderPreservingPartitioners. it may be better to create > another key CF for sorting the ke

Re: [cassandra] sorting TimeUUID column?

setReversed(true) sorts timeUUID in descending order. thanks, -aj On Sun, May 9, 2010 at 6:44 AM, Lee Parker wrote: > You need to set the Reversed attribute on the sliceRange. I don't know > which library you are using, but it may be something like > sliceRange.setReversed(true). > > Lee Parke

Re: key is sorted?

true, i think. but don't rush to OrderPreservingPartitioners. it may be better to create another key CF for sorting the key (put the key in column, which is sorted automatically). to paginate through the records, you get a sorted slice of keys from the key CF, and then look up the record CF with ea

Re: Extremly slow inserts on LAN

From a naive (not caring about Cassandra internals) basis, the first step is to isolate whether the problem is on the client or server side. Have you tried a Linux-based server or a Linux-based client? On 2010-05-09 14:06, Arie Keren wrote: > While making our first steps with Cassandra, we experie

Extremly slow inserts on LAN

While making our first steps with Cassandra, we experience slow inserts working on LAN. Inserting 7000 keys with 1 column family takes about 10 seconds when Cassandra server running on the same host with the client. But when server runs on a different host on LAN, the same inserts take more th

Re: release date for 0.7 ?

Here are a few things for 0.7: http://svn.apache.org/repos/asf/cassandra/trunk/NEWS.txt Not sure about a date. Eben On Sat, May 8, 2010 at 8:57 PM, vineet daniel wrote: > Hi > > What is the expected release date for 0.7 and what will be t

Re: [cassandra] sorting TimeUUID column?

You need to set the Reversed attribute on the sliceRange. I don't know which library you are using, but it may be something like sliceRange.setReversed(true). Lee Parker On Sat, May 8, 2010 at 7:36 PM, AJ Chen wrote: > I have a column family for sorting time-uuid and expect to retrieve columns

Re: key is sorted?

Thanks! I just wanted to be extra sure about that On Sun, May 9, 2010 at 3:54 PM, Shinpei Ohtani wrote: > > But you will get all the keys, right? > > Yes. > > On Sun, May 9, 2010 at 9:45 PM, David Boxenhorn wrote: > > But you will get all the keys, right? > > > > I want to get all the keys,

Re: key is sorted?

> But you will get all the keys, right? Yes. On Sun, May 9, 2010 at 9:45 PM, David Boxenhorn wrote: > But you will get all the keys, right? > > I want to get all the keys, but I don't care about the order. > > On Sun, May 9, 2010 at 3:25 PM, Shinpei Ohtani > wrote: >> >> Hi, >> >> > If you use

Re: key is sorted?

But you will get all the keys, right? I want to get all the keys, but I don't care about the order. On Sun, May 9, 2010 at 3:25 PM, Shinpei Ohtani wrote: > Hi, > > > If you use random partitioning, you can loop through all keys with a > range > > query, but they will not be sorted. > > > > True

Re: key is sorted?

Hi, > If you use random partitioning, you can loop through all keys with a range > query, but they will not be sorted. > > True or False? I think they will be sorted in a way of MD5 hash if you use RandomPartitioner, at least I looked into SSTables and RandomPartitioner#getToken. On Sun, May 9,

Re: key is sorted?

This is something that I'm not sure that I understand. Can somebody confirm/deny that I understand it? Thanks. If you use random partitioning, you can loop through all keys with a range query, but they will not be sorted. True or False? On Sat, May 8, 2010 at 3:45 AM, AJ Chen wrote: > thanks,