Question about JSR 326

[Ankit bhatt <an...@elitecore.com>]

Hello,

 

I have gone through JCP site for understanding of JSR 326, as well I have
also gone through Kato site. I am bit confused about the main purpose for
this JSR. As far as I can understand, there are APIs available for Profiling
in Java, like HPROF and JPDA, but they are capable enough to track JVM when
it's running and the purpose can be like Performance tuning or JVM class
loading tracing etc.

 

So JSR 326 will be used in the situations when JVM crashed due to some
Errors, is it? It will provide API support for tracking such kind of Events
as well will provide details about JVM state at that point, is it?

 

I will be grateful to you, if I can get the answers from your side.

 

Thanking You. 

 

Thanks & Regards,

Ankit Bhatt.

 

 

Re: Question about JSR 326

[Steve Poole <sp...@googlemail.com>]

Hi  - thanks for taking the trouble to ask us directly.   Looks like we need
to do a better job of communicating the objectives of the API !

Now to answer your questions.


On Sat, Oct 3, 2009 at 8:11 AM, Ankit bhatt <an...@elitecore.com>wrote:

> Hello,
>
>
>
> I have gone through JCP site for understanding of JSR 326, as well I have
> also gone through Kato site. I am bit confused about the main purpose for
> this JSR. As far as I can understand, there are APIs available for
> Profiling
> in Java, like HPROF and JPDA, but they are capable enough to track JVM when
> it's running and the purpose can be like Performance tuning or JVM class
> loading tracing etc.
>
>
>
> So JSR 326 will be used in the situations when JVM crashed due to some
> Errors, is it? It will provide API support for tracking such kind of Events
> as well will provide details about JVM state at that point, is it?
>
>
>
You are  almost correct - at the highest level  the API is simply intended
to be an abstraction layer over diagnostics artifacts (ie dumps). That means
that if you have a dump from a JVM (such as HPROF or a corefile etc)  you
will be able to view the data in a standard form.   Dumps are normally
associated with failing applications or JVMs but there are reasons why that
is not always true.  Dumps contain a great deal of useful information that
can help with monitoring live systems.   So this API is not just about
helping you understand crashed JVMs - it will be used as part of trend
analysis.

The API has two basic layers at the moment and may gain a third one before
completion.  The  first layer presents information about a process (such as
environmental variables, loaded libraries,  native threads and access to
process memory info )    The second layer provides an abstraction of a Java
runtime which means you can find out about things like  loaded classes, Java
threads, GC root information and what objects are on the heap.


Hope that answers your questions -  please ask again if its not clear.



>
> I will be grateful to you, if I can get the answers from your side.
>
>
>
> Thanking You.
>
>
>
> Thanks & Regards,
>
> Ankit Bhatt.
>
>
>
>
>
>


-- 
Steve