Closed GoogleCodeExporter closed 9 years ago
Hi,
Thanks for reporting a bug.
After you edited eclipse.ini to run Eclipse with JDK1.6(C:\Program
Files\java\jdk1.6.0_22) instead of JRE1.6 (C:\Program Files\Java\jre6), does
the issue still occur when changing the debug launch configuration to
"C:\program files (x86)\java\jdk1.6.0_18"?
Best Regards,
Yoshitaka
Original comment by yoshitak...@gmail.com
on 1 Dec 2010 at 1:02
I could not reproduce the issue in the following environment that is the same
as reported.
OS: Windows 7 64bit
Eclipse: 3.6.1
JVM Monitor: 3.6.6
JVM for Eclipse: JRE 1.6.0_22
JVM for monitored Java application: JDK 1.6.0_18 or JDK 1.6.0_22
JVM Monitor shows the message "Monitoring is not supported for the selected
JVM" if the target JVM cannot be connected by invoking the SUN JDK API
com.sun.tools.attach.VirtualMachine.attach(). For example, when trying to
connect to JVM running with JRE 1.5, the exception
com.sun.tools.attach.AttachNotSupportedException is thrown and the message "..
not supported.." is shown. The message doesn't always mean a bug, but there is
certainly a case that the JVM is not supported to be connected depending on
environment.
If there is no additional feedback, this entry will be closed as
not-reproducible.
Original comment by yoshitak...@gmail.com
on 5 Dec 2010 at 11:29
The followings are the typical cases that JVM Monitor shows the message
"Monitoring is not supported for the selected JVM" on Properties view on
Windows 64bit.
case 1) running Java app with Java 1.5
case 2) running Java app with JRE/JDK 32bit while using Eclipse 64bit on Windows 64bit
case 3) running Java app with JRE/JDK 64bit while using Eclipse 32bit on Windows 64bit
This bug entry seems to be the case 2, and it is expected behavior to show
error message. On the other hand, the existing message is not that descriptive,
so it would make sense to show additional message contained in exception
com.sun.tools.attach.AttachNotSupportedException
on Properties view and status line when selecting the JVM on JVM Explorer as
below.
additional message to be shown:
case 1) The VM does not support the attach mechanism
case 2) Unable to attach to 32-bit process running under WOW64
case 3) Unable to attach to 64-bit process
Original comment by yoshitak...@gmail.com
on 14 Dec 2010 at 10:17
I am seeing this issue on Windows 64-bit using the latest versions of Eclipse
(SpringSource Tool Suite 2.7, based on Eclipse Indigo) and JVM Monitor
(3.7.0.201106191853). No additional message is shown; however, the error "Could
not open PerfMemory" appears in the Eclipse error log (see attached).
As indicated in the attached Eclipse configuration, I am starting Eclipse using
JDK 1.6.0_23 located in C:/Program Files/Java/jdk1.6.0_23/bin. If I start my
Java application from a command prompt window using this path to the java
executable explicitly (e.g., "C:\Program Files\Java\jdk1.6.0_23\bin\java ..."
the PID shows up in JVM Explorer and I get the above error when selecting it.
The Eclipse instance also appears in JVM Explorer and I am able to attach to it.
This may be a different issue from the original poster as I think I am using
the same VM for both Eclipse and the application I want to profile.
Original comment by maf9676...@gmail.com
on 5 Jul 2011 at 9:11
Attachments:
Follow-up on the above: found a discussion related to the PerfMemory error at
http://forums.oracle.com/forums/thread.jspa?messageID=7067232. This discussion
indicated that this can be due to a JVM bug related to inherited permissions
and gave a work-around of changing the TMP environment variable to reference
C:\TMP (which must exist) for both the monitoring and monitored applications.
Doing so worked for me; perhaps the code can identify this error and provide
the workaround?
Original comment by maf9676...@gmail.com
on 5 Jul 2011 at 10:01
Hi.
Just to say the workaround about TMP variable is a mandatory FAQ entry.
Every time I use JVMMONITOR with a new computer, I had to digg into my memory
or google to found out this solution.
Thanks
Original comment by marcde...@gmail.com
on 24 Sep 2012 at 3:49
Original issue reported on code.google.com by
thegar...@gmail.com
on 30 Nov 2010 at 4:45Attachments: