dtrace discuss - Jan 2008 - Truning on pid$target probes hangs Solaris

Hi:

I''m running Indiana (OpenSolaris developer preview) under parallels
in OS X 10.5.1. I can consistently get Solaris to hang.

Doing a
dtrace -n pid\$target::malloc:entry''{@=count()}'' -c ls

will consistently hang Solaris.

Basically doing anything with PID provider with the
-c option seems to hangs Solaris. Only way to recover
is to stop and start Parallels desktop.

The same D one liner on a running process works just fine. (ie)
dtrace -n pid1234::malloc:entry''{@=count()}'' has no problem.

I tried to look at Parallels during the hang using pid provider in
OS X and I see it is doing a bunch of ioctl''s on /dev/vm-main

 From within Solaris if I start dtrace under truss, then I find the last
thing truss was able to record is a bunch of preads. I have the
output from truss if anyone is interested. (its 180K)

This could not have happened at the worse time. I was before
an important ISV showing off DTrace and how stable it is and
it hangs on me!

Any help on debugging this would be great.

Thanks

Angelo

I''d suggest trying to connect with the Parallels folks to see if they
can
tell you where we are in the Solaris kernel.

I''d also suggest that VMware Fusion might be a better virtualization
option.
That''s not to say we haven''t has issues with VMware, but they
have been very
helpful in diagnosing problems and helping us solve them:

  http://blogs.sun.com/ahl/entry/debugging_cross_calls_on_opensolaris

Adam

On Wed, Jan 30, 2008 at 05:06:08PM -0500, Angelo Rajadurai
wrote:
> Hi:
> 
> I''m running Indiana (OpenSolaris developer preview) under
parallels
> in OS X 10.5.1. I can consistently get Solaris to hang.
> 
> Doing a
> dtrace -n pid\$target::malloc:entry''{@=count()}'' -c ls
> 
> will consistently hang Solaris.
> 
> Basically doing anything with PID provider with the
> -c option seems to hangs Solaris. Only way to recover
> is to stop and start Parallels desktop.
> 
> The same D one liner on a running process works just fine. (ie)
> dtrace -n pid1234::malloc:entry''{@=count()}'' has no
problem.
> 
> I tried to look at Parallels during the hang using pid provider in
> OS X and I see it is doing a bunch of ioctl''s on /dev/vm-main
> 
>  From within Solaris if I start dtrace under truss, then I find the last
> thing truss was able to record is a bunch of preads. I have the
> output from truss if anyone is interested. (its 180K)
> 
> This could not have happened at the worse time. I was before
> an important ISV showing off DTrace and how stable it is and
> it hangs on me!
> 
> Any help on debugging this would be great.
> 
> Thanks
> 
> Angelo
> 
> _______________________________________________
> dtrace-discuss mailing list
> dtrace-discuss at opensolaris.org
-- 
Adam Leventhal, Fishworks                        http://blogs.sun.com/ahl

Hi Adam:

Thanks. I can confirm that this does not happen on VMWare fusion 1.1.1
Will work with the parallels folks to see if they can provide some  
more info
on what the kernel is doing

Thanks

Angelo


On Jan 30, 2008, at 5:28 PM, Adam Leventhal wrote:
> I''d suggest trying to connect with the Parallels folks to see if  
> they can
> tell you where we are in the Solaris kernel.
>
> I''d also suggest that VMware Fusion might be a better
virtualization
> option.
> That''s not to say we haven''t has issues with VMware, but
they have
> been very
> helpful in diagnosing problems and helping us solve them:
>
> http://blogs.sun.com/ahl/entry/debugging_cross_calls_on_opensolaris
>
> Adam
>
> On Wed, Jan 30, 2008 at 05:06:08PM -0500, Angelo Rajadurai wrote:
>> Hi:
>>
>> I''m running Indiana (OpenSolaris developer preview) under
parallels
>> in OS X 10.5.1. I can consistently get Solaris to hang.
>>
>> Doing a
>> dtrace -n pid\$target::malloc:entry''{@=count()}'' -c
ls
>>
>> will consistently hang Solaris.
>>
>> Basically doing anything with PID provider with the
>> -c option seems to hangs Solaris. Only way to recover
>> is to stop and start Parallels desktop.
>>
>> The same D one liner on a running process works just fine. (ie)
>> dtrace -n pid1234::malloc:entry''{@=count()}'' has no
problem.
>>
>> I tried to look at Parallels during the hang using pid provider in
>> OS X and I see it is doing a bunch of ioctl''s on /dev/vm-main
>>
>> From within Solaris if I start dtrace under truss, then I find the  
>> last
>> thing truss was able to record is a bunch of preads. I have the
>> output from truss if anyone is interested. (its 180K)
>>
>> This could not have happened at the worse time. I was before
>> an important ISV showing off DTrace and how stable it is and
>> it hangs on me!
>>
>> Any help on debugging this would be great.
>>
>> Thanks
>>
>> Angelo
>>
>> _______________________________________________
>> dtrace-discuss mailing list
>> dtrace-discuss at opensolaris.org
>
> -- 
> Adam Leventhal, Fishworks                        http://blogs.sun.com/ahl