Hi Bill,
Kohsuke and I too looked into btrace.
Infact, Kohsuke and I had a discussion with Sundar too after J1 and have
metnioned some of our gprobes ideas. He also agreed that there needs to
be some changes to btrace implementation to meet some of our needs. We
can certainly use some of the ideas OR use it as an underlying farmework
(after changing it) for glassfish.
Regarding dtrace support in JDK7,
The approach mentioned in the blog
(
http://blogs.sun.com/kamg/entry/announcing_statically_defined_dtrace_probes)
requires a static probe provider interface. This is exactly same as what
we are planning to do in gprobes for glassfish. We can use the same
interface (after dynamically retransforming this interface to extend
com.sun.tracing.Provider) to obtain a dtrace ProbeProvider.
Thanks,
--Mahesh
Sahoo wrote:
> I had briefly looked at btrace sometime back. It's a very useful
> serviceability tool. Sundar, the author of btrace, is ready to help us
> if we want to use it in GlassFish. I know he was writing some probes
> for the JDBC layer of GlassFish to demonstrate the power of btrace
> during btrace BoF at J1 this year.
>
> Thanks,
> Sahoo
>
> Bill Shannon wrote:
>> Has anyone played with btrace (https://btrace.dev.java.net/)?
>>
>> Has anyone looked at adding dtrace probes to GlassFish?
>> http://blogs.sun.com/kamg/entry/announcing_statically_defined_dtrace_probes
>>
>> Hopefully they'll make the com.sun.tracing package available for
>> pre-JDK7 releases, at least as a nop, so we can start putting in
>> probes now.
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe_at_glassfish.dev.java.net
>> For additional commands, e-mail: dev-help_at_glassfish.dev.java.net
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: dev-help_at_glassfish.dev.java.net
>