I''m new to DTrace; been working through the "Solaris Performance and Tools" book this morning. DTrace is working great, I''ve learned some interesting things, and have run a number of the DTrace Toolkit scripts and written a few tiny ones. However, I can''t get any of the hotspot probes to work. DTrace reports no probes that match hotspot. I''m running the OpenSolaris 11/08 release, using the default java (which is 1.6.10). See the following: root at opensolaris:/usr/Emporos# dtrace -l | wc 56144 279580 4262222 root at opensolaris:/usr/Emporos# dtrace -ln ''hotspot*:::'' ID PROVIDER MODULE FUNCTION NAME dtrace: failed to match hotspot*:::: No probe matches description root at opensolaris:/usr/Emporos# java -version java version "1.6.0_10" Java(TM) Platform, Standard Edition for Business (build 1.6.0_10-b33) Java HotSpot(TM) Server VM (build 11.0-b15, mixed mode) -- This message posted from opensolaris.org
I''ve seen this message when I''m in a zone... works in global zone, in my case... Russ Ted Graham wrote:> ... > dtrace: failed to match hotspot*:::: No probe matches description > >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://mail.opensolaris.org/pipermail/dtrace-discuss/attachments/20081215/a149c847/attachment.html>
If you do an ''elfdump'' on the <JAVA_HOME>/jre/lib/i386/client/libjvm.so file in the JDK (well, any libjvm.so you find in there will work), does it show any SUNW_dof sections? If it doesn''t, then I suspect the binary in that package was built incorrectly. If you download the JDK from java.sun.com and install it manually, do you get the same problem? I suspect the java.sun.com package might work. If so, then there''s a problem with the Java package that''s being distributed for opensolaris. -- - Keith Ted Graham wrote:> I''m new to DTrace; been working through the "Solaris Performance and Tools" book this morning. DTrace is working great, I''ve learned some interesting things, and have run a number of the DTrace Toolkit scripts and written a few tiny ones. > > However, I can''t get any of the hotspot probes to work. DTrace reports no probes that match hotspot. I''m running the OpenSolaris 11/08 release, using the default java (which is 1.6.10). > > See the following: > > root at opensolaris:/usr/Emporos# dtrace -l | wc > 56144 279580 4262222 > > root at opensolaris:/usr/Emporos# dtrace -ln ''hotspot*:::'' > ID PROVIDER MODULE FUNCTION NAME > dtrace: failed to match hotspot*:::: No probe matches description > > root at opensolaris:/usr/Emporos# java -version > java version "1.6.0_10" > Java(TM) Platform, Standard Edition for Business (build 1.6.0_10-b33) > Java HotSpot(TM) Server VM (build 11.0-b15, mixed mode)
Ted, First, welcome to DTrace: I think you''ll enjoy the ride. Since the hotspot provider is a USDT (User Statically Defined Tracing) provider, the provider name is going to be hotspot1234, 1234 is the PID of the JVM process. The probes don''t get created until the JVM is running, and they go away when the JVM exits. And just in case this question comes up, some of probes that tend to fire very often require the -XX:+ExtendedDTraceProbes JVM option to be active. In you haven''t seen it, see http://java.sun.com/javase/6/docs/technotes/guides/vm/dtrace.html for what I believe is the official documentation for the hotspot probes. Keep in mind that even though the documentation show the "args" array being used, USDT probes don''t support the args array. You have to use the built-in variables, arg0, arg1, etc. Chip> -----Original Message----- > From: dtrace-discuss-bounces at opensolaris.org [mailto:dtrace-discuss- > bounces at opensolaris.org] On Behalf Of Ted Graham > Sent: Monday, December 15, 2008 11:22 AM > To: dtrace-discuss at opensolaris.org > Subject: [dtrace-discuss] No hotspot probes available? > > I''m new to DTrace; been working through the "Solaris Performance and > Tools" book this morning. DTrace is working great, I''ve learned some > interesting things, and have run a number of the DTrace Toolkitscripts> and written a few tiny ones. > > However, I can''t get any of the hotspot probes to work. DTracereports> no probes that match hotspot. I''m running the OpenSolaris 11/08 > release, using the default java (which is 1.6.10). > > See the following: > > root at opensolaris:/usr/Emporos# dtrace -l | wc > 56144 279580 4262222 > > root at opensolaris:/usr/Emporos# dtrace -ln ''hotspot*:::'' > ID PROVIDER MODULE FUNCTION > NAME > dtrace: failed to match hotspot*:::: No probe matches description > > root at opensolaris:/usr/Emporos# java -version > java version "1.6.0_10" > Java(TM) Platform, Standard Edition for Business (build 1.6.0_10-b33) > Java HotSpot(TM) Server VM (build 11.0-b15, mixed mode) > -- > This message posted from opensolaris.org > _______________________________________________ > dtrace-discuss mailing list > dtrace-discuss at opensolaris.org
Thanks for the quick responses! It appears that you are correct, the binary
delivered with OpenSolaris 11/08 doesn''t have the hotspot provider.
Once I installed a new Java from Sun, I have the hotspot probes. Thanks much,
it would have been a long time before I figured that out.
Ted
See the elfdump output below (there is no SUNW_dof), hopefully what is there
means something to you.
admin at opensolaris:~$ elfdump
/usr/jdk/instances/jdk1.6.0/jre/lib/i386/libjvm.so | grep SUNW
p_paddr: 0 p_type: [ PT_SUNWCAP ]
Section Header[1]: sh_name: .SUNW_cap
sh_size: 0x10 sh_type: [ SHT_SUNW_cap ]
[248] 0x00000000 0x00000000 OBJT GLOB D 0 ABS
SUNWprivate_1.1
[2] 0x000000b4 0x00000000 SECT LOCL D 0 .SUNW_cap
[29] 0x00000000 0x00000000 SECT LOCL D 0 .SUNW_cap
[31] 0x00000000 0x00000000 SECT LOCL D 0 .SUNW_cap
[4691] 0x00000000 0x00000000 OBJT LOCL D 0 .SUNW_cap _START_
[18750] 0x00000000 0x00000000 OBJT GLOB D 0 ABS
SUNWprivate_1.1
[248] SUNWprivate_1.1
[25] SUNW_CAP 0xb4
Hardware/Software Capabilities Section: .SUNW_cap
[0] CA_SUNW_HW_1 0x5 [ CX8 FPU ]
--
This message posted from opensolaris.org
I''m not familiar with opensolaris packaging. We should somehow get a bug report into them about this. Is anyone on this list familiar with the process or able to do that? -- - Keith Ted Graham wrote:> Thanks for the quick responses! It appears that you are correct, the binary delivered with OpenSolaris 11/08 doesn''t have the hotspot provider. Once I installed a new Java from Sun, I have the hotspot probes. Thanks much, it would have been a long time before I figured that out. > > Ted > > See the elfdump output below (there is no SUNW_dof), hopefully what is there means something to you. > > admin at opensolaris:~$ elfdump /usr/jdk/instances/jdk1.6.0/jre/lib/i386/libjvm.so | grep SUNW > p_paddr: 0 p_type: [ PT_SUNWCAP ] > Section Header[1]: sh_name: .SUNW_cap > sh_size: 0x10 sh_type: [ SHT_SUNW_cap ] > [248] 0x00000000 0x00000000 OBJT GLOB D 0 ABS SUNWprivate_1.1 > [2] 0x000000b4 0x00000000 SECT LOCL D 0 .SUNW_cap > [29] 0x00000000 0x00000000 SECT LOCL D 0 .SUNW_cap > [31] 0x00000000 0x00000000 SECT LOCL D 0 .SUNW_cap > [4691] 0x00000000 0x00000000 OBJT LOCL D 0 .SUNW_cap _START_ > [18750] 0x00000000 0x00000000 OBJT GLOB D 0 ABS SUNWprivate_1.1 > [248] SUNWprivate_1.1 > [25] SUNW_CAP 0xb4 > Hardware/Software Capabilities Section: .SUNW_cap > [0] CA_SUNW_HW_1 0x5 [ CX8 FPU ]
Anyone can report a bug against OpenSolaris at http://defect.opensolaris.org/ -Alan Coopersmith- alan.coopersmith at sun.com Sun Microsystems, Inc. - X Window System Engineering Keith McGuigan wrote:> I''m not familiar with opensolaris packaging. We should somehow get a > bug report into them about this. Is anyone on this list familiar with > the process or able to do that? > > -- > - Keith > > Ted Graham wrote: >> Thanks for the quick responses! It appears that you are correct, the binary delivered with OpenSolaris 11/08 doesn''t have the hotspot provider. Once I installed a new Java from Sun, I have the hotspot probes. Thanks much, it would have been a long time before I figured that out. >> >> Ted >> >> See the elfdump output below (there is no SUNW_dof), hopefully what is there means something to you. >> >> admin at opensolaris:~$ elfdump /usr/jdk/instances/jdk1.6.0/jre/lib/i386/libjvm.so | grep SUNW >> p_paddr: 0 p_type: [ PT_SUNWCAP ] >> Section Header[1]: sh_name: .SUNW_cap >> sh_size: 0x10 sh_type: [ SHT_SUNW_cap ] >> [248] 0x00000000 0x00000000 OBJT GLOB D 0 ABS SUNWprivate_1.1 >> [2] 0x000000b4 0x00000000 SECT LOCL D 0 .SUNW_cap >> [29] 0x00000000 0x00000000 SECT LOCL D 0 .SUNW_cap >> [31] 0x00000000 0x00000000 SECT LOCL D 0 .SUNW_cap >> [4691] 0x00000000 0x00000000 OBJT LOCL D 0 .SUNW_cap _START_ >> [18750] 0x00000000 0x00000000 OBJT GLOB D 0 ABS SUNWprivate_1.1 >> [248] SUNWprivate_1.1 >> [25] SUNW_CAP 0xb4 >> Hardware/Software Capabilities Section: .SUNW_cap >> [0] CA_SUNW_HW_1 0x5 [ CX8 FPU ] > _______________________________________________ > dtrace-discuss mailing list > dtrace-discuss at opensolaris.org