gmane.comp.java.netbeans.modules.profiler.user

Call Sequence Tracing with profiler

Hi list,

I've spent a day or so investigating the possibility to trace Java  
method invocation sequences (to understand behaviors of an existing  
application upon some user actions) .

Docs show Eclipse TPTP has some ability to generate Sequence Diagram  
in UML as profiling, but I've never succeeded with TPTP even to start  
profiling.

The closest result I achieved so far is to view a CPU snapshot in NB  
profiler by sorting with the "Call Tree" column, however that won't  
show the exact order of method invocations in time sequence.

I'd downloaded a source zip of whole NB and digged a little into  
profiler, my first idea had been to add a "Time Occurred" or "Sequence  
No." column to have the CCT sortable by occurrence sequence, but found  
no data field to back it. And I see the CCT nodes are *physically*  
reordered on sorting, and there's a default sort, so the *natural*  
order of CCT nodes is never available.

Am I expecting the wrong thing from profiler? Or how is "Call Sequence  
Tracing" possible with profiler (regarding Eclipse TPTP seems to  
promise that) ?

Any idea appreciated, thanks in advance!
-
Compl

headers

Dennis Weyland | 16 Sep 14:22

Problem getting the profiler working

Hello!

I have a problem getting the profiler working. I hangs during the 
calibration process at "connection to the target vm". I tried out three 
different (of the currently with my packet manager available) VMs 
(java-6-sun-1.6.0.06, java-6-openjdk, java-1.5.0-sun-1.5.0.15). I am 
using Netbeans 6.0.1 on Ubuntu.

Thank you for your help.

Best regards,
   Dennis

headers

Albertas Janulevicius - TNW | 6 Aug 22:01

Profiling CPU time

Hi guys, is it normal that the time spent in a particular method (as shown in a call tree) depends on the filter used?

headers

Albertas Janulevicius - TNW | 17 Jul 15:54

Different profiling results

Hi,

I am new to NetBeans Profiler, so sorry if I ask a stupid question. I performed profiling of "Entire application" with several different filters, and looked at the total time spend in a particular method. I would expect the total time to be the same, but for each filter I get a different result. Furthermore, if a method is a bottleneck with one filter, with another filter it becomes much faster and some other method becomes the slowest one. Just want to stress that I am talking about the total time, not about the self-time (I understand that the self time can change depending on the filter). I tried to switch off the dynamic frequency scaling in BIOS, but still have the same problem.

Could not find anything about this on the net... Could anyone please help me to understand what's going on?

headers

Andrew Novoselsky | 17 Jul 09:34

CVM crash in case of JIT enabled and attached NetBeansProfiler

HI,
Now we are working on porting NetBeans Profiler on Linux MIPS platform with CDC. We fully suceeded with CDC 72 revision. But on CDC rev 74 we faced with CVM crash in following case:
1) CDC is built with JIT enabled
2) Start our application with attached profiler
3) Connect from Netbeans IDE to CVM
4) Profiler instrument classes
5) crash

Execution fails in src\share\javavm\runtime\constantpool.c file in CVMprivate_cpExtractTypeIDFromUnresolvedEntry function at line 294 :
typeIDIdx = CVMcpGetMemberRefTypeIDIdx(cp, cpIndex);
Crash is caused by assertation macros :
+(CVMassert(CVMcpTypeIs(cp, idx, Fieldref) \+
+|| CVMcpTypeIs(cp, idx, Methodref) \+
|| CVMcpTypeIs(cp, idx, InterfaceMethodref))
as argument tipe is incorrect (not equals to Fieldref or Methodref or InterfaceMethodref)

Can anyone suggest any workarounds ?

--
Best wishes, Andrew V. Novoselsky

headers

Aaron Kaplan | 9 Jul 18:19

Redefinition errors from memory profiler

When I try to run my project in the memory profiler, I get a dozen
dialog boxes with messages of the form

  Redefinition failed with error xx
  Check JVMTI documentation for this error code

where xx is 63, 64, 66, or 67.  Then I get the following fatal error:

Exception in thread "main" java.lang.NoSuchMethodError:
org.apache.commons.logging.impl.WeakHashtable$Referenced.<init>(Lorg/apache/commons/logging/impl/WeakHashtable$2;Ljava/lang/Object;)V

The redefinition errors are all for classes in the packages
com.hp.hpl.jena and org.apache.commons.logging.

What do I do?
Thanks
-Aaron

headers

Jaroslav Bachorik | 30 Apr 17:57

[Fwd: [nbj2ee] Cannot open temporary communication file created at the server side]

forwarding to the appropriate mailing list

From: Sedillo, Derek <derek.sedillo@...>
Subject: [nbj2ee] Cannot open temporary communication file created at the server side
Date: 2008-04-30 15:49:36 GMT

Hello,

I am using the NetBeans Profiler for a Java application running remotely
on a Solaris system.  I got everything setup for the remote connection.
And the profiler works in Thread mode.  But when I switch to memory mode
it errors out.  Here is the message from the log on the Windows client:

IDEAppStatusHandler - error: Cannot open temporary communication file
created at the server side.
Name:  

* Notice the name field is blank

On the Solaris system I have made the home directory read/write world.
So writing shouldn't be a problem.

I saw this question posted a while back with no response.  Does anyone
have information of this.  It may be a bug with the profiler remote
functionality.

Thank you,

Derek Sedillo
Security Product Engineering 
v622-5234 (719-535-5234)

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@...
For additional commands, e-mail: users-help@...

headers

刘Q | 20 Apr 04:18

0企业信息

1ÆÚ´ýÄúµÄÀ´µç£¡

   ÏêÇéÇë¿´¸½¼þ

Attachment (ͼƬ1.JPG): application/octet-stream, 148 KiB

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@...
For additional commands, e-mail: users-help@...

headers

Marcus Meyerhöfer | 9 Apr 16:21

Profiling Tomcat - Out of Memory

Dear Netbeans Profiler Community!

I am trying to profile memory of a Tomcat 5.5.23 using JDK 1.5.0_12 on a Windows XP system. I modified the
catatina.bat as described and am able to start the Tomcat instance (which has the application I am
interested in deployed as usual) which waits until the profiler connects. However,
after some time of progress where live profiling results are displayed the Tomcat instance either stops
responding (which results in the Netbeans profiler asking wether to resume profiling or not) or I receive
a " Profiler Agent Communication Thread OutOfMemoryError" in the cmd-shell i used to start
catalina_nbprofiler.bat. I assume that the first case (Tomcat not responding) has the same cause as this
behaviour always occurs when the heap size in the VM Telemetry view reaches about 80M.

The webapp IŽm trying to profile is started with -Xms382m and -Xmx1024m. 

So my question is wether there are some parameters or configuration options to grant more memory to the
profiler agent? I couldnŽt find any information on this issue...

Thanks for any advice.

Greetings

Marcus
_________________________________________________________________________
In 5 Schritten zur eigenen Homepage. Jetzt Domain sichern und gestalten! 
Nur 3,99 EUR/Monat! http://www.maildomain.web.de/?mc=021114

headers

Michael Chernyshov | 26 Mar 16:47

Profiler support for JavaFX Script plugin: go to source issue

Hi team,

we're working on JavaFX profiler and have a problem with "go to source" 
feature reported today by Martin Ryzl (thanks, Martin!): 
http://javafx.netbeans.org/issues/show_bug.cgi?id=131143.

Our team is highly interested in the SPI mentioned in the issue to be 
implemented ASAP since we'd like to have a full functional profiler for 
our plugin.

If you need more information from our team or have some suggestions what 
we could do to enhance JavaFX profiler functionality, we'd appreciate 
your input since we're not familiar with profiler code.

If you'd like to try the JavaFX Script plugin to look into JavaFX 
profiler in its current state, you can use the daily builds: 
http://deadlock.netbeans.org/hudson/job/Java_FX_NB_plugin_Trunk_daily

Thanks in advance,
Mike.

headers

Jess Holle | 26 Mar 12:44

Profiler crash bug...

It looks like NetBeans bug #100641 
[http://www.netbeans.org/issues/show_bug.cgi?id=100641] should be 
re-opened -- based JDC bug #6546266 
[http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6546266].

If so VisualVM definitely needs this particular fix.  Having the 
profiler crash whenever you attempt to profile Tomcat is pretty hard to 
explain to a development team.  They tend to simply decide the whole 
tool is untrustworthy.  Plus there's the fact that profiling Tomcat is 
actually quite necessary at times...  [Tomcat is just as effected by 
this issue as JBoss.]

--
Jess Holle

Mon	Tue	Wed	Thu	Fri	Sat	Sun

		1	2	3	4	5
6	7	8	9	10	11	12
13	14	15	16	17	18	19
20	21	22	23	24	25	26
27	28	29	30	31