Re: How to implement (generic) ACID on application level

... some kind of what? On Mon, Sep 6, 2010 at 3:38 AM, Michal Augustýn wrote: > Thank you for the great link! > The mentioned solution is using locking but I would prefer some optimistic > strategy (because the conflicts are rare in my situation) but I'm afraid > that this is really the best solu

Re: busy thread on IncomingStreamReader ?

I encounter the same problem, does anyone have sovled the problem? On Tue, Apr 20, 2010 at 11:03 AM, Ingram Chen wrote: > I check system.log both, but there is no exception logged. > > On Tue, Apr 20, 2010 at 10:40, Jonathan Ellis wrote: > >> I don't see csArena-tmp-6-Index.db in the incoming

Flush and compaction happens frequently

I have a 8 nodes cluster, MemtableThreshold is 2 GB/CF, MemtableObjectsCount is 1.2, heap (min/max) is 30 GB and only 4 ColumnFamilies It appears from system.log that flush happens for every < 50 operations (read or write) and compaction is happening very frequently and i could see lots of sstable

Re: 4k keyspaces... Maybe we're doing it wrong?

On Mon, Sep 6, 2010 at 12:41 AM, Janne Jalkanen wrote: > > So if I read this right, using lots of CF's is also a Bad Idea(tm)? > Yes, lots of CFs is bad means lots of CFs is also bad.

Re: Migration from 6.X to 7.X

On Mon, Sep 6, 2010 at 4:04 PM, Edward Capriolo wrote: > I was not aware of that. Also is the default for 6.o non framed and > 7.o framed? Yes. >I was thinking possibly replace cassanda.client detect the > server version and use reflection. This would have to be tested; the speed penalty for us

Re: Migration from 6.X to 7.X

Welcome to Thrift. On Mon, Sep 6, 2010 at 4:04 PM, Edward Capriolo wrote: > > I was not aware of that. Also is the default for 6.o non framed and > 7.o framed? I was thinking possibly replace cassanda.client detect the > server version and use reflection. This way hector sees the same > interface

Re: Migration from 6.X to 7.X

On Monday, September 6, 2010, Jonathan Ellis wrote: > Thrift does not support method overloading (methods with the same name > but different parameters). > > On Mon, Sep 6, 2010 at 9:09 AM, Edward Capriolo wrote: >> On Mon, Sep 6, 2010 at 3:33 AM, Ran Tavory wrote: >>> we don't have one version

Re: Migration from 6.X to 7.X

Thrift does not support method overloading (methods with the same name but different parameters). On Mon, Sep 6, 2010 at 9:09 AM, Edward Capriolo wrote: > On Mon, Sep 6, 2010 at 3:33 AM, Ran Tavory wrote: >> we don't have one version that supports both versions. >> you can hack it if you downloa

Re: Migration from 6.X to 7.X

On Mon, Sep 6, 2010 at 3:33 AM, Ran Tavory wrote: > we don't have one version that supports both versions. > you can hack it if you download the source code (create two java package > trees for 0.6.0 and 0.7.0) but it's not on the shelf, sorry... > > On Mon, Sep 6, 2010 at 12:39 AM, Edward Capriol

Re: skip + limit support in GetSlice

Thank you! This solve my issue. But what about index recomputing (after new columns are inserted) ? Should I use asynchronous triggers? https://issues.apache.org/jira/browse/CASSANDRA-1311 Or will 0.7's secondary indexes handle this? Augi 20

Re: How to implement (generic) ACID on application level

Thank you for the great link! The mentioned solution is using locking but I would prefer some optimistic strategy (because the conflicts are rare in my situation) but I'm afraid that this is really the best solution... So the solution is probably to use some kind of 2010/9/6 Reza Lesmana > I re

RE: skip + limit support in GetSlice

Have you considered creating a second column family which acts as an index for the original column family? Have the record number as the column name, and the value as the identifier (primary key) of the original data, and do a 1. get_slice(, start='00051235', finish='', limit=100) 2. get_slic

Re: skip + limit support in GetSlice

Hi Mike, yes, I read the PDF to the finish. Twice. As I wrote, my application is not accessed by users, it's accessed by other applications that can access pages randomly. So when some application wants to get page 51235 (so skip is 5123500, limit is 100) then I have to: 1) GetSlice(from: "", to

Re: 4k keyspaces... Maybe we're doing it wrong?

#2. Performance: Will Cassandra work better with a single keyspace + lots of keys, or thousands of keyspaces? Thousands is a non-starter. There is an active memtable for every CF defined and caches (row and key) are per CF. Assuming even 2 CFs per keyspace, with 4000 keyspaces you will hav