I track a very hard memory leak (or most likely left memory). I do the following:
1) start my application
2) get to the point where the application will show a leak
3) start the tools using the "distribution" option
4) join my process and start recording
5) take the source heap / 6) take the initial snapshot using VM tracker
7 ) repeat the steps that make memory grow 8) take another heap 9) take another snapshot using VM tracker
If I take these steps, I see results that really don't make sense. I hope that I do not have enough information about how these tools work. For example, I know that the leak tool does not track all kinds of memory allocations (like carbon applications). My application is a huge, inherited application that may have a strange distribution code in some outdated subsystem that I am not familiar with. However, this is what I see:
- in the initial snapshot of VM tracking, when I look at the resume, the type MALLOC_SMALL is pretty small at about 72 MB (virtual size).
- in the second snapshot, the use of MALLOC_SMALL has grown to 224 MB (virtual size again)
- in the second snapshot, he tells me that the heap of memory is 45 MB.
, MALLOC_SMALL 72 224 , 45 ? Allocations -, VMTracker?
, - Allocations... , MALLOC_SMALL (, , ), 72 - > 224 , ? , (, 0x79000000-0x7b000000) Allocations " " . . , 4 , 4 ?! 32 , VM- ?
... , - , , .