dtrace discuss - Mar 2007 - Chime systemic unresponsiveness

I was viewing the two dtrace screen cast presentations that Eric mentioned in:
http://www.opensolaris.org/jive/thread.jspa?threadID=25604&tstart=0
by Brendan and Tom which are booth very good and now we expect all presentations
to be done this way now. :-)

 In Tom Ericksons Chime he was tracing syscalls with 
chime -n ''syscall:::entry { @[execname] = count();}''
and on my AMD Athlon64 dual core I get an "Abort due to systemic
unresponsiveness "  with the java backtace as:

DTrace consumer started at Mar 6, 2007 12:18:23 AM PST. Error message received
at Mar 6, 2007 12:18:33 AM PST: Abort due to systemic unresponsiveness  Chime
Stack Trace: org.opensolaris.os.dtrace.DTraceException: Abort due to systemic
unresponsiveness         org.opensolaris.os.dtrace.LocalConsumer._consume(Native
Method)        
org.opensolaris.os.dtrace.LocalConsumer.work(LocalConsumer.java:624)        
org.opensolaris.os.dtrace.LocalConsumer$1.run(LocalConsumer.java:707)        
java.lang.Thread.run(Thread.java:619)

 Taking one core offline with psradm fixes the problem.  Build 57, I''m
assuming java build 1.6.0-b105 was the default.   Any idea what could be causing
this?

---Bob
 
 
This message posted from opensolaris.org

> In Tom Ericksons Chime he was tracing syscalls with
>  
> chime -n ''syscall:::entry { @[execname] = count();}''
> and on my AMD Athlon64 dual core I get an "Abort due
> to systemic unresponsiveness " 
I think this is Bug ID 6507659, "tsc differences between CPU''s
give dtrace_gethrtime() serious problems":
http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6507659

Fixed in snv_58.
 
 
This message posted from opensolaris.org

> > In Tom Ericksons Chime he was tracing syscalls
> with
> >  
> > chime -n ''syscall:::entry { @[execname] >
count();}''
> > and on my AMD Athlon64 dual core I get an "Abort
> due
> > to systemic unresponsiveness " 
> 
> I think this is Bug ID 6507659, "tsc differences
> between CPU''s
> give dtrace_gethrtime() serious problems":
> http://bugs.opensolaris.org/bugdatabase/view_bug.do?bu
> g_id=6507659
> 
> Fixed in snv_58.
 Ah yes it works in snv_59, thanks.

---Bob
 
 
This message posted from opensolaris.org