what is this “predictive self-healing” for solaris 10? is it marketing speak for automatic update?
or is it a “safety net” for crashing apps?
predictive? which means it learns? learns that a particular app is always calling illegal memory and freezing it? learning that a paricular user-supplied kernel module alwaysleads to a crash? or does it modify code/data on the fly? how does it know what is best?
what is this “predictive self-healing” for solaris 10? is it marketing speak for automatic update?
You’ve correctly identified this as marketing-speak. However, it turns out that Predictive Self Healing is actually comprised of SMF and FMA, which are the Service Management Framework and the Fault Management Architecture, respectively.
Probably one of the best features of this is automatic detection and isolation of bad hardware. In previous version of Solaris, if a CPU goes bad, often the kernel would simply crash with “BAD TRAP” errors or something to that effect and you’d be left with a kernel core file you’d have to send to Sun’s kernel group to be analyzed, at which point they could tell you what hardware is dying. Predictive self-healing would automatically notice a bad CPU and shut it down. In most Sun hardware it could then be hot swapped for a new one without even having to power the system down.
Sun’s Predictive Self-Healing (PSH) architecture changes the approach used for model for error handling and fault management used in a traditional UNIX operating system. The architecture moves from an error centric model where haphazard error messages are generated and left for humans to diangose, isolate and recover from the underlying problem to a model where we can rapidly deploy software to do the same tasks with greater efficiency and accuracy. Where possible, the Solaris 10 diagnosis technology will predict when a component is at risk for failure and gracefully retire it from the system until a repair can be scheduled.
PSH also provides a new adminstrative model for system and application services. In addition to a comprehensive management model for services, the Service Management Facility (SMF) provides tighter dependency controls between system resources (CPUs, memory, and IO) and application services. The coupling of system resources and services permits Solaris to safely recover (self-heal) services in the face of hardware and software faults while providing clear instructions on how to repair the problem.
So. what is this sudden influx on solaris related articles.
Uh… prob because the new version is going to be shipped to customers in general release next month.
And also because SysadminMag has a regular (annual)
solaris issue. Comes out in December every year.
what is this “predictive self-healing” for solaris 10? is it marketing speak for automatic update?
or is it a “safety net” for crashing apps?
predictive? which means it learns? learns that a particular app is always calling illegal memory and freezing it? learning that a paricular user-supplied kernel module alwaysleads to a crash? or does it modify code/data on the fly? how does it know what is best?
what is this “predictive self-healing” for solaris 10? is it marketing speak for automatic update?
You’ve correctly identified this as marketing-speak. However, it turns out that Predictive Self Healing is actually comprised of SMF and FMA, which are the Service Management Framework and the Fault Management Architecture, respectively.
There’s a 10,000 ft. overview of these here:
http://iforce.sun.com/protected/solaris10/adoptionkit/general/featu…
If that’s not available for some reason, a more general summary is here:
http://www.sun.com/bigadmin/content/selfheal/
Probably one of the best features of this is automatic detection and isolation of bad hardware. In previous version of Solaris, if a CPU goes bad, often the kernel would simply crash with “BAD TRAP” errors or something to that effect and you’d be left with a kernel core file you’d have to send to Sun’s kernel group to be analyzed, at which point they could tell you what hardware is dying. Predictive self-healing would automatically notice a bad CPU and shut it down. In most Sun hardware it could then be hot swapped for a new one without even having to power the system down.
So. what is this sudden influx on solaris related articles.
It’s “SysAdmin Magazine”… got it?
“It’s “SysAdmin Magazine”… got it?”
no. poor reasoning
Sun’s Predictive Self-Healing (PSH) architecture changes the approach used for model for error handling and fault management used in a traditional UNIX operating system. The architecture moves from an error centric model where haphazard error messages are generated and left for humans to diangose, isolate and recover from the underlying problem to a model where we can rapidly deploy software to do the same tasks with greater efficiency and accuracy. Where possible, the Solaris 10 diagnosis technology will predict when a component is at risk for failure and gracefully retire it from the system until a repair can be scheduled.
PSH also provides a new adminstrative model for system and application services. In addition to a comprehensive management model for services, the Service Management Facility (SMF) provides tighter dependency controls between system resources (CPUs, memory, and IO) and application services. The coupling of system resources and services permits Solaris to safely recover (self-heal) services in the face of hardware and software faults while providing clear instructions on how to repair the problem.
More details on PSH may be found at:
http://www.sun.com/bigadmin/content/selfheal
Or follow the discussion forum:
http://forum.sun.com/forum.jspa?forumID=241