1

Topic: About big

In respect of an exchange of experience: than use for the analysis large ?  so from 80. I.e. when storage of a normal desktop does not suffice even for indexes. The purposes simple enough, normally it is necessary to clarify what objects led OOM to understand what not so happened in system. I while use here it  NetBeans profiler and I write the code to pull out interesting things. But it would be desirable any more declarative way, like language of requests.

2

Re: About big

kl> I while use here it  NetBeans profiler and I write the code to pull out interesting things. But it would be desirable any more declarative way, like language of requests. Is better to adjust histogram printing in a broad gull at OOM - it sufficed me.

3

Re: About big

Hello, bzig, you wrote: B> Is better to adjust histogram printing in a broad gull at OOM - it sufficed me. I and from   will always construct the Histogram without UI is not a problem. Interesting as it is convenient to write requests for object graph bypass in big  to find leaks etc. And that of the histogram I simply I see that, say, class  contains 100500  in lump in 50, but in system they can form in a heap of places and it is desirable to have the additional information (who holds these  up to GC root etc.).

4

Re: About big

kl> (who holds these  up to GC root etc.). Among HZ  will be and , so to you and so it is necessary to reconsider the code. kl> but in system they can form in a heap of places the Heap is how many? 10-20-50 is all it is possible to check up in an hour-two, armed .