What is causing this error: dtrace: processing aborted: Abort due to systemic unresponsiveness It happens to my script which shows all the NFS operations coming through common_dispatch() on a server. I wasn''t the one running the test, so I don''t know how long it ran before it died. The script uses a couple of static probes and a couple fbt. Thanks, jim -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://mail.opensolaris.org/pipermail/dtrace-discuss/attachments/20051117/47fa8adf/attachment.html>
James C. McPherson
2005-Nov-17 22:12 UTC
[dtrace-discuss] Abort due to systemic unresponsiveness
Jim Wahlig wrote:> What is causing this error: > dtrace: processing aborted: Abort due to systemic unresponsiveness > It happens to my script which shows all the NFS operations coming > through common_dispatch() on a server. > I wasn''t the one running the test, so I don''t know how long it ran > before it died. > The script uses a couple of static probes and a couple fbt.Jim, your script is causing DTrace to hit its internal safety valves. I''ve seen this when following an interrupt on my x64 motherboard and while causing an interrupt storm :| Was your server busy? James C. McPherson -- Solaris Datapath Engineering Data Management Group Sun Microsystems
Jonathan Adams
2005-Nov-17 22:21 UTC
[dtrace-discuss] Abort due to systemic unresponsiveness
On Thu, Nov 17, 2005 at 03:55:39PM -0600, Jim Wahlig wrote:> What is causing this error: > > dtrace: processing aborted: Abort due to systemic unresponsivenessThis is the dtrace deadman timer; if dtrace thinks that it may be causing the system to become unresponsive, it will turn itself off. The criteria is: - an interrupt can fire once a second - The dtrace(1M) consumer can run once every thirty seconds You can turn it off by enabling destructive actions (add a -w). Cheers, - jonathan -- Jonathan Adams, Solaris Kernel Development
Bryan Cantrill
2005-Nov-17 22:29 UTC
[dtrace-discuss] Abort due to systemic unresponsiveness
On Fri, Nov 18, 2005 at 09:12:39AM +1100, James C. McPherson wrote:> Jim Wahlig wrote: > >What is causing this error: > >dtrace: processing aborted: Abort due to systemic unresponsiveness > >It happens to my script which shows all the NFS operations coming > >through common_dispatch() on a server. > >I wasn''t the one running the test, so I don''t know how long it ran > >before it died. > >The script uses a couple of static probes and a couple fbt. > > Jim, > your script is causing DTrace to hit its internal safety valves. > I''ve seen this when following an interrupt on my x64 motherboard > and while causing an interrupt storm :|If this is a busy NFS server, it''s possible that it''s _so_ busy that dtrace(1M) can''t run. If dtrace(1M) can''t run for 40 seconds, dtrace(7D) will get nervous and abort the enabling. If you believe that the machine is otherwise functional, you can try running dtrace(1M) in the RT scheduling class: # priocntl -e -c RT dtrace -s ./myscript.d ... This will allow dtrace(1M) to preempt your in-kernel NFS activity. If this still fails, then you have a very, very busy machine -- and if you still believe that this is in error, you can enable destructive actions (thereby disabling the deadman). But (and this should go without saying) be very careful doing this... - Bryan -------------------------------------------------------------------------- Bryan Cantrill, Solaris Kernel Development. http://blogs.sun.com/bmc
Seemingly Similar Threads
- [Fwd: Persistent "Abort due to systemic unresponsiveness"]
- plockstat: processing aborted: Abort due to systemic unresponsiveness
- dtrace: processing aborted: Abort due to systemic unresponsiveness
- problem with smbios 2.13?
- Functions that cannot be traced by FBT under x86