Hi Ruslan,

Is it possible for you to tell us the details on what you have done which measurably helped your situation, so we can start a "best practices" doc on growing cassandra systems? 

So far, I see that under load, cassandra is rarely "ready" to take heavy load in it's default configuration and a number of steps need to be done with the configuration of cassandra for proper sizing of memtables, flushing, jvm.  Unfortunately, it's very difficult to gauge what the proper or appropriate settings are for a given workload.

It would be helpful if you could share, what happened in the default config, what steps you did that helped the situation, h Tow much each step helped your situation.  That way we can start a checklist of things to address as we grow in load.

Paul

On 3/8/2011 5:15 AM, ruslan usifov wrote:


2011/3/8 Paul Pak <ppak@yellowseo.com>
Hi Ruslan,

It looks like Jonathan and Stu have already been working to reduce garbage collection on v.8  The ticket is at https://issues.apache.org/jira/browse/CASSANDRA-2252

Jonathan, is there any way to apply the patch to .73 and have ruslan test it to see if it fixes his issue with Garbage collection?  Given the fact that he's already got a test case, it seems like a reasonable thing to do (with the caveat that he is a guinea pig for new code which may not be fully tested yet :) .

Thanks Paul for information, it very informative. But how this work at others. I see that situation stay better when i increase heap size to 7GB, but nodes still frozen truth not so frequent as before, but still, also but I don;t probe this i think that setting -Xmn to something bigger then 800M may give benefit (http://java.sun.com/performance/reference/whitepapers/tuning.html)