Registered: Aug 2000
Location:
Posts: 1725

So I went to a Solaris 10 "Boot Camp" today put on by Bob Netherton of SUN. What a good show. I have already played with quite a few of the subjects that he spoke of, such a ZFS and Zones, and I have a little experience using DTRACE instead of truss... but his examples and utilization of the tool took it to a new height.

How great would it be, to actually see how many sub-routines one process spawns, and exactly see the syscall functions they run away process calls on? Can you say... "Never gonna have a doubt about whats gone awry!", or "Never going to have to spend 4 hours second guessing". DTRACE eliminates that factor by allowing you to see whats really gone wrong. You can literally trace stuff down deep within the depths of the system level to observe in real time the issues you may be having with a run away process that NO OTHER TOOLS can show you. You can actually attach dtrace to a specific PID and watch it go... altering the view by utilizing the D language. Without going into a long long long discussion over the great stuff DTRACE can do, Ill just point you to this site: http://blogs.sun.com/bobn

Other stuff he went over is Predictive Self-Healing, which is pretty much self explainitory, and VERY cool.

I highly suggest installing Solaris 10, or Solaris Express and playing with all the innovations that SUN has made...

Report this post to a moderator | IP: Logged