Contributing Factors
Poorly written software items such as:
1. De-allocating pointers to allocated memory and continuing to use the pointer
2. Case statements with no default case
3. State machines that can erroneously jump state
4. Incomplete error handlers
These type of bugs have been found in application stack, low level drivers and are often found when using net back up and Veritas software.
Cases involving third party PCI cards not operating to PCI 1.2 specifications. These are timing problems where the PCI card is sending a signal to the JBUS at the wrong time, resulting in an error.
Symptoms
If this issue occurs the following error will be seen;
Fatal Error Reset CPU 0000.0000.0000.0003 AFSR 0200.0000.0000.0000 JETO
AFAR 0000.07c6.0000.0f00 Executing Power On Self Test
When a customer experiences a JETO gather all the data and escalate to TSC as described in the Resolution section below.
Root Cause
A JETO is the result of the CPU seeking data across the JBUS and when the data is not delivered to the requesting CPU within the allocated time, a JBUS time out results. This JETO causes a fatal system reset, where no state is saved and the registers do not have any meaningful data, so there is no known way to fix this problem. The host system resets to avoid a hang system
Corrective Action
Resolution
*DO NOT REPLACE CPU, MEMORY, OR MOTHER BOARD until instructed to by TSC.*
Swapping hardware will not resolve this JETO error.
Collect the following information:
1.Please make sure all the ODM steps are completed.
Information on the Oracle Diagnostic Methodology (ODM) can be found via the below URL;
https://support.us.oracle.com/oip/faces/secure/km/DocumentDisplay.jspx?id=312789.1
2. How many systems are seeing this issue at this site?
3. Determine what the system configuration is, including PCI cards and which slots the cards are installed into?
4. What applications were running at the time of the fatal? Please note if Veritas backup software was in use?
5. What storage devices are attached to this system?
6. Verify the PCI configuration to determine if we can try an alternate layout of cards.
7. Determine the real impact on the Customer's operation, the software and the application.
8. Review the patch levels of the Customers COMPLETE software stack including Veritas and other components
9.Gather an explorer.
10. Obtain system history.
If all cards are Sun supported and the systems configuration looks good, TSC will explain to the customer that they are to monitor the system, and if they get another JETO to call into the center with all the data and reference the prior case ID.
TSC will work with Engineering to resolve known issues. If the Engineering actions fail, TSC will provide you with an Escalation path.
Identification of Affected Parts (how to):
There is no affected hardware. Report the explorer file to TSC.
Comments:
These issues are very difficult to resolve so it is highly recommended that you work with TSC.
Modification History:
15-May-2014: Maintenance check for relevance/currency, no change in content
For information about FABs you may email:
