 |
 |
|
| Home > | Next > |
| | | | | |||||||||||||||||||||||||||||||||||||||||||
|
|
|
Fast Track to Solaris 10 Adoption: Predictive Self-HealingGeneral InformationPlease click on a question below or download a pdf version.
Q: How about General Overview of Predictive Self-Healing? A: Sun has developed a new architecture for building and deploying systems and services capable of Predictive Self-Healing. Self-healing technology enables Sun systems and services to maximize availability in the face of software and hardware faults and facilitates a simpler and more effective end-to-end experience for system administrators, thus reducing cost of ownership. Q: How long have you been planning this release? A: Some of the technology inside the Solaris 10 OS, has been under development for up to three years, a little bit longer for the design and architecture. It's been a long, yet gratifying road. Q: How can I get the software? A: There are two flavors of the Solaris Express program. If you want to receive support, you get the "Commercial Use" version and pay $99. If you just wish to begin testing the Solaris 10 OS, we recommend the free version. You can get it here. Things might be a bit more clear. You can get CD images for either SPARC or x86 platforms, Commercial or Non-Commercial use. Q: What are some of the new features in Solaris 10 OS? A: Some of the new features you can find in Solaris 10 OS are DTrace, Predictive Self-Healing, Solaris Grid Containers and Dynamic File Systems. For more information please check http://www.sun.com/solaris and click on the Solaris 10 link. A: Yes, we are working on making sure that there is a good connection with the remote facilities. Q: What is different between the Solaris 9 and 10 operating systems? A: There are many new features inside the Solaris 10 OS that you don't find in the Solaris 9 OS, such as DTrace, Predictive Self-Healing, and Solaris Grid Containers. You can read more about it at http://www.sun.com/solaris/10. Remember, there is binary compatibility between Solaris 9 and 10 operating systems. A: In the short term, our immediate goals are to port everything we've done so far to x86/AMD. Going forward, we want to both build a more complete profile of "system-level" objects (e.g., harden more drivers), then move up the stack to higher level abstractions (e.g. file systems, network, Java). Q: So SMF is not in Beta 6, but will it be in Beta 7? A: Yes, it is scheduled to be in Beta 7. A: Service Management Facility. A: We're actively exploring connections to higher-level network management software. Stay tuned to PSH announcements for more information. A: There are plans to instrument SVM and our future file systems for PSH. Migration of tools such as mdlogd will occur at that time. Q: How does PSH relate to SNMP? A: SNMP is a way to produce traps that monitoring software use, and Predictive Self-Healing is a framework inside the operating system that trigger actions and corrective procedures, you can find more information at http://www.sun.com/msg Q: Will this be included in the Solaris open source distribution? A: Yes, Sun's intent is to make Solaris open source, and that would include the source for PSH features. Q: Can you briefly discuss the idea surrounding telemetry probes and how you reached that concept? A: In traditional Unix systems, error information is haphazardly output to syslog for humans to diagnose the underlying fault. The error messages are confusing and difficult to write software to interpret and respond to. With PSH, we have moved away from unstable error messages to a new protocol for describing error information, using stable, self-describing data. The error data forms the basis of the new error stream (telemetry) that can be dispatched to software for programmatic diagnosis. A: For software services, the next release of Solaris Express will include the second part of PSH, which is the Service Management Facility. This performs monitoring of software applications and uses PSH to know when to automatically restart them. You can look at Stephen Hahn's blog (blogs.sun.com/sch) for a preview. For your memory question, we will recognize faults within the memory space of any user process. If the fault causes correctable errors (CEs), we'll diagnose the issue and potentially re-map the physical page under the process. If the process hits an uncorrectable error (UE) in memory, it will be killed, then SMF will restart the containing service. So PSH addresses faults that affect the kernel and those that affect user processes. A: Application developers will be allowed to participate as an SMF service that may be restarted in the face of a fault on one or more resources it is using. Resource abstractions via Java will continue to work but may not be fully connected to SMF resource service contracts. Q: What will be the FCS date for the Solaris 10 OS? A: The Solaris 10 OS will ship at the end of 2004. Q: Will PSH have a connection to Netconnect? A: This is part of the roadmap for PSH, but it will not be available in the first release of the Solaris 10 OS. A: Yes, it is available for both platforms through http://www.sun.com/software/solaris/solaris-express/ Q: Can PSH be installed on top of the Solaris 9 OS? A: That is a new feature of the Solaris 10 OS. Q: Write to future delivery of PSH — are plans to expand the functionality outside of systems? A: Predictive Self-Healing is a new fault management technology being implemented across Sun's product line. It is a Sun-wide effort. Q: How is this different from the Solaris 9 OS? A: Predictive Self-Healing is not part of the Solaris 9 OS. It is a new technology in the Solaris 10 OS.
| | |||||||||||||||||||||||||||||||||||||||||||
