Jim Fiori
2010-Aug-13 13:45 UTC
[dtrace-discuss] Why does java not stop() when I tell it to stop()?
Hi,
I''m using the rexit/stop()/prun() dance to capture process exit in
order to grab various procfs commands on exit. Works fine for simple program
like "sleep". But for Java (and other threaded customer apps) it runs
one or 2 of the p-commands, but the process goes away before I can get all I
want:
$ cat exit.d
#!/usr/sbin/dtrace -qws
syscall::rexit:entry
/execname == "sleep" || execname == "java"/
{
stop();
system("pargs %d",pid);
system ("pldd %d",pid);
system ("pfiles %d",pid);
system ("pstack %d",pid);
system ("prun %d",pid);
}
$ ./exit.d&
[1] 22487
OK, now test a "sleep" exit:
bash-3.00$ sleep 5
22488: sleep 5
argv[0]: sleep
argv[1]: 5
22488: sleep 5
/lib/libc.so.1
/usr/lib/locale/en_US.ISO8859-1/en_US.ISO8859-1.so.3
22488: sleep 5
Current rlimit: 256 file descriptors
0: S_IFCHR mode:0620 dev:287,0 ino:12582952 uid:5667 gid:7 rdev:24,18
O_RDWR
/devices/pseudo/pts at 0:18
1: S_IFCHR mode:0620 dev:287,0 ino:12582952 uid:5667 gid:7 rdev:24,18
O_RDWR
/devices/pseudo/pts at 0:18
2: S_IFCHR mode:0620 dev:287,0 ino:12582952 uid:5667 gid:7 rdev:24,18
O_RDWR
/devices/pseudo/pts at 0:18
22488: sleep 5
feea4d18 ecvt (2, 8047290, 8047296, 0, 8047298, 8047306)
Looks like all the p-commands completed. But for java:
0$ /usr/java/bin/java -version
java version "1.5.0_20"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.5.0_20-b02)
Java HotSpot(TM) Client VM (build 1.5.0_20-b02, mixed mode, sharing)
22513: /usr/java/bin/java -version
argv[0]: /usr/java/bin/java
argv[1]: -version
22513: /usr/java/bin/java -version
/lib/libthread.so.1
/lib/libdl.so.1
/lib/libc.so.1
/usr/jdk/instances/jdk1.5.0/jre/lib/i386/client/libjvm.so
/lib/libsocket.so.1
/usr/lib/libsched.so.1
/usr/lib/libCrun.so.1
/lib/libm.so.1
/lib/libnsl.so.1
/lib/libm.so.2
/lib/libscf.so.1
/lib/libdoor.so.1
/lib/libuutil.so.1
/lib/libgen.so.1
/lib/libmd.so.1
/lib/libmp.so.2
/usr/jdk/instances/jdk1.5.0/jre/lib/i386/native_threads/libhpi.so
/usr/jdk/instances/jdk1.5.0/jre/lib/i386/libverify.so
/usr/jdk/instances/jdk1.5.0/jre/lib/i386/libjava.so
/usr/jdk/instances/jdk1.5.0/jre/lib/i386/libzip.so
/usr/lib/locale/en_US.ISO8859-1/en_US.ISO8859-1.so.3
bash-3.00$ pfiles: cannot examine 22513: no such process
pstack: cannot examine 22513: no such process or core file
prun: cannot control 22513: no such process
Any ideas? Something is telling the process to continue.
Jim
--
This message posted from opensolaris.org
Adam Leventhal
2010-Aug-13 18:56 UTC
[dtrace-discuss] Why does java not stop() when I tell it to stop()?
Hey Jim, I don''t have any ideas off the top of my head, but you can certainly use DTrace to figure out what''s giving java the order to continue... Adam On Aug 13, 2010, at 6:45 AM, Jim Fiori wrote:> Hi, > > I''m using the rexit/stop()/prun() dance to capture process exit in order to grab various procfs commands on exit. Works fine for simple program like "sleep". But for Java (and other threaded customer apps) it runs one or 2 of the p-commands, but the process goes away before I can get all I want: > > $ cat exit.d > #!/usr/sbin/dtrace -qws > > syscall::rexit:entry > /execname == "sleep" || execname == "java"/ > { > stop(); > system("pargs %d",pid); > system ("pldd %d",pid); > system ("pfiles %d",pid); > system ("pstack %d",pid); > > system ("prun %d",pid); > } > > $ ./exit.d& > [1] 22487 > > OK, now test a "sleep" exit: > > bash-3.00$ sleep 5 > 22488: sleep 5 > argv[0]: sleep > argv[1]: 5 > 22488: sleep 5 > /lib/libc.so.1 > /usr/lib/locale/en_US.ISO8859-1/en_US.ISO8859-1.so.3 > 22488: sleep 5 > Current rlimit: 256 file descriptors > 0: S_IFCHR mode:0620 dev:287,0 ino:12582952 uid:5667 gid:7 rdev:24,18 > O_RDWR > /devices/pseudo/pts at 0:18 > 1: S_IFCHR mode:0620 dev:287,0 ino:12582952 uid:5667 gid:7 rdev:24,18 > O_RDWR > /devices/pseudo/pts at 0:18 > 2: S_IFCHR mode:0620 dev:287,0 ino:12582952 uid:5667 gid:7 rdev:24,18 > O_RDWR > /devices/pseudo/pts at 0:18 > 22488: sleep 5 > feea4d18 ecvt (2, 8047290, 8047296, 0, 8047298, 8047306) > > Looks like all the p-commands completed. But for java: > > 0$ /usr/java/bin/java -version > java version "1.5.0_20" > Java(TM) 2 Runtime Environment, Standard Edition (build 1.5.0_20-b02) > Java HotSpot(TM) Client VM (build 1.5.0_20-b02, mixed mode, sharing) > 22513: /usr/java/bin/java -version > argv[0]: /usr/java/bin/java > argv[1]: -version > 22513: /usr/java/bin/java -version > /lib/libthread.so.1 > /lib/libdl.so.1 > /lib/libc.so.1 > /usr/jdk/instances/jdk1.5.0/jre/lib/i386/client/libjvm.so > /lib/libsocket.so.1 > /usr/lib/libsched.so.1 > /usr/lib/libCrun.so.1 > /lib/libm.so.1 > /lib/libnsl.so.1 > /lib/libm.so.2 > /lib/libscf.so.1 > /lib/libdoor.so.1 > /lib/libuutil.so.1 > /lib/libgen.so.1 > /lib/libmd.so.1 > /lib/libmp.so.2 > /usr/jdk/instances/jdk1.5.0/jre/lib/i386/native_threads/libhpi.so > /usr/jdk/instances/jdk1.5.0/jre/lib/i386/libverify.so > /usr/jdk/instances/jdk1.5.0/jre/lib/i386/libjava.so > /usr/jdk/instances/jdk1.5.0/jre/lib/i386/libzip.so > /usr/lib/locale/en_US.ISO8859-1/en_US.ISO8859-1.so.3 > bash-3.00$ pfiles: cannot examine 22513: no such process > pstack: cannot examine 22513: no such process or core file > prun: cannot control 22513: no such process > > > Any ideas? Something is telling the process to continue. > > Jim > -- > This message posted from opensolaris.org > _______________________________________________ > dtrace-discuss mailing list > dtrace-discuss at opensolaris.org-- Adam Leventhal, Fishworks http://blogs.sun.com/ahl
Jim Fiori
2010-Aug-13 23:01 UTC
[dtrace-discuss] Why does java not stop() when I tell it to stop()?
Thanks Adam - I''ve already started down this path. Jim> Hey Jim, > > I don''t have any ideas off the top of my head, but > you can certainly use DTrace to figure out what''s > giving java the order to continue... > > Adam > > On Aug 13, 2010, at 6:45 AM, Jim Fiori wrote: > > > Hi, > > > > I''m using the rexit/stop()/prun() dance to capture > process exit in order to grab various procfs commands > on exit. Works fine for simple program like "sleep". > But for Java (and other threaded customer apps) it > runs one or 2 of the p-commands, but the process goes > away before I can get all I want: > > > > $ cat exit.d > > #!/usr/sbin/dtrace -qws > > > > syscall::rexit:entry > > /execname == "sleep" || execname == "java"/ > > { > > stop(); > > system("pargs %d",pid); > > system ("pldd %d",pid); > > system ("pfiles %d",pid); > > system ("pstack %d",pid); > > > > system ("prun %d",pid); > > } > > > > $ ./exit.d& > > [1] 22487 > > > > OK, now test a "sleep" exit: > > > > bash-3.00$ sleep 5 > > 22488: sleep 5 > > argv[0]: sleep > > argv[1]: 5 > > 22488: sleep 5 > > /lib/libc.so.1 > > > /usr/lib/locale/en_US.ISO8859-1/en_US.ISO8859-1.so.3 > > 22488: sleep 5 > > Current rlimit: 256 file descriptors > > 0: S_IFCHR mode:0620 dev:287,0 ino:12582952 > uid:5667 gid:7 rdev:24,18 > > O_RDWR > > /devices/pseudo/pts at 0:18 > > 1: S_IFCHR mode:0620 dev:287,0 ino:12582952 > uid:5667 gid:7 rdev:24,18 > > O_RDWR > > /devices/pseudo/pts at 0:18 > > 2: S_IFCHR mode:0620 dev:287,0 ino:12582952 > uid:5667 gid:7 rdev:24,18 > > O_RDWR > > /devices/pseudo/pts at 0:18 > > 22488: sleep 5 > > feea4d18 ecvt (2, 8047290, 8047296, 0, 8047298, > 8047306) > > > > Looks like all the p-commands completed. But for > java: > > > > 0$ /usr/java/bin/java -version > > java version "1.5.0_20" > > Java(TM) 2 Runtime Environment, Standard Edition > (build 1.5.0_20-b02) > > Java HotSpot(TM) Client VM (build 1.5.0_20-b02, > mixed mode, sharing) > > 22513: /usr/java/bin/java -version > > argv[0]: /usr/java/bin/java > > argv[1]: -version > > 22513: /usr/java/bin/java -version > > /lib/libthread.so.1 > > /lib/libdl.so.1 > > /lib/libc.so.1 > > > /usr/jdk/instances/jdk1.5.0/jre/lib/i386/client/libjvm > .so > > /lib/libsocket.so.1 > > /usr/lib/libsched.so.1 > > /usr/lib/libCrun.so.1 > > /lib/libm.so.1 > > /lib/libnsl.so.1 > > /lib/libm.so.2 > > /lib/libscf.so.1 > > /lib/libdoor.so.1 > > /lib/libuutil.so.1 > > /lib/libgen.so.1 > > /lib/libmd.so.1 > > /lib/libmp.so.2 > > > /usr/jdk/instances/jdk1.5.0/jre/lib/i386/native_thread > s/libhpi.so > > > /usr/jdk/instances/jdk1.5.0/jre/lib/i386/libverify.so > > /usr/jdk/instances/jdk1.5.0/jre/lib/i386/libjava.so > > /usr/jdk/instances/jdk1.5.0/jre/lib/i386/libzip.so > > > /usr/lib/locale/en_US.ISO8859-1/en_US.ISO8859-1.so.3 > > bash-3.00$ pfiles: cannot examine 22513: no such > process > > pstack: cannot examine 22513: no such process or > core file > > prun: cannot control 22513: no such process > > > > > > Any ideas? Something is telling the process to > continue. > > > > Jim > > -- > > This message posted from opensolaris.org > > _______________________________________________ > > dtrace-discuss mailing list > > dtrace-discuss at opensolaris.org > > > -- > Adam Leventhal, Fishworks > http://blogs.sun.com/ahl > _________________________ > dtrace-discuss mailing list > dtrace-discuss at opensolaris.org >-- This message posted from opensolaris.org
Jim Fiori
2010-Aug-16 16:54 UTC
[dtrace-discuss] Why does java not stop() when I tell it to stop()?
Well it turns out that various proc commands (pldd, pstack to name a few)
send process control messages to the process by writing to /proc/PID/ctl.
One of those commands is PCRUN which makes the process runnable.
I noticed though that if I stopped the process outside of DTrace with pstop,
the process stayed stopped even after running these procfs commands.
So all I had to do to work around this was add a pstop command after the DTrace
stop():
stop();
system("pstop %d",pid)
... collect till my heart''s content
system("prun %d",pid);
Jim
> Hi,
>
> I''m using the rexit/stop()/prun() dance to capture
> process exit in order to grab various procfs commands
> on exit. Works fine for simple program like "sleep".
> But for Java (and other threaded customer apps) it
> runs one or 2 of the p-commands, but the process goes
> away before I can get all I want:
>
> $ cat exit.d
> #!/usr/sbin/dtrace -qws
>
> syscall::rexit:entry
> /execname == "sleep" || execname == "java"/
> {
> stop();
> system("pargs %d",pid);
> system ("pldd %d",pid);
> system ("pfiles %d",pid);
> system ("pstack %d",pid);
> system ("prun %d",pid);
>
>
> $ ./exit.d&
> [1] 22487
>
> OK, now test a "sleep" exit:
>
> bash-3.00$ sleep 5
> 22488: sleep 5
> argv[0]: sleep
> argv[1]: 5
> 22488: sleep 5
> /lib/libc.so.1
> /usr/lib/locale/en_US.ISO8859-1/en_US.ISO8859-1.so.3
> 22488: sleep 5
> Current rlimit: 256 file descriptors
> 0: S_IFCHR mode:0620 dev:287,0 ino:12582952 uid:5667
> gid:7 rdev:24,18
> O_RDWR
> /devices/pseudo/pts at 0:18
> 1: S_IFCHR mode:0620 dev:287,0 ino:12582952
> uid:5667 gid:7 rdev:24,18
> O_RDWR
> /devices/pseudo/pts at 0:18
> 2: S_IFCHR mode:0620 dev:287,0 ino:12582952
> uid:5667 gid:7 rdev:24,18
> O_RDWR
> /devices/pseudo/pts at 0:18
> 2488: sleep 5
> feea4d18 ecvt (2, 8047290, 8047296, 0, 8047298,
> 8047306)
>
> Looks like all the p-commands completed. But for
> java:
>
> 0$ /usr/java/bin/java -version
> java version "1.5.0_20"
> Java(TM) 2 Runtime Environment, Standard Edition
> (build 1.5.0_20-b02)
> Java HotSpot(TM) Client VM (build 1.5.0_20-b02, mixed
> mode, sharing)
> 22513: /usr/java/bin/java -version
> argv[0]: /usr/java/bin/java
> argv[1]: -version
> 22513: /usr/java/bin/java -version
> /lib/libthread.so.1
> /lib/libdl.so.1
> /lib/libc.so.1
> /usr/jdk/instances/jdk1.5.0/jre/lib/i386/client/libjvm
> .so
> /lib/libsocket.so.1
> /usr/lib/libsched.so.1
> /usr/lib/libCrun.so.1
> /lib/libm.so.1
> /lib/libnsl.so.1
> /lib/libm.so.2
> /lib/libscf.so.1
> /lib/libdoor.so.1
> /lib/libuutil.so.1
> /lib/libgen.so.1
> /lib/libmd.so.1
> /lib/libmp.so.2
> /usr/jdk/instances/jdk1.5.0/jre/lib/i386/native_thread
> s/libhpi.so
> /usr/jdk/instances/jdk1.5.0/jre/lib/i386/libverify.so
> /usr/jdk/instances/jdk1.5.0/jre/lib/i386/libjava.so
> /usr/jdk/instances/jdk1.5.0/jre/lib/i386/libzip.so
> /usr/lib/locale/en_US.ISO8859-1/en_US.ISO8859-1.so.3
> bash-3.00$ pfiles: cannot examine 22513: no such
> process
> pstack: cannot examine 22513: no such process or core
> file
> prun: cannot control 22513: no such process
>
>
> Any ideas? Something is telling the process to
> continue.
>
> Jim
--
This message posted from opensolaris.org