If you encounter a situation similar to this issue using a different I/O card, please contact Sun Support Services with the failure details and the CR and this SRDB will need to be updated with regards to the new failure symptoms/information.

Below is the process which will confirm that this document and CR 6238924, applies to your failure situation. It shows how to reproduce the issue and also how to test the workarounds. If the issue can not be reproduced or the workarounds do not resolve your specific situation, then this document and CR may not apply to your situation.

Please make sure that this process is executed in the domain console and that console logging is enabled so this information can be sent in to Sun Support Services for documentation purposes.

Please validate that each troubleshooting step below is true for your environment. Each step will provide instructions via a link to the document, for validating the step and taking corrective action as necessary. The steps are ordered in the most appropriate sequence to isolate the issue and identify the proper resolution. Please do not skip a step.

Process:

1) I/O card is located in I/O Board slot 0 (I/O Board number is not relevant). Card scsi port 0 is attached to scsi device (assumes Cauldron-S - and the scsi device attached is a D240 array).

   - Again, not sure if Cauldron-S is the only card susceptible to this;  
     But, at this time it is the only reported adapter to have this issue.

2) Setkeyswitch on.

3) At OBP set diag-switch to true.

   - This will show "probing for device" information as the domain exits POST
     and enters OBP.  This is useful trouble shooting information if needed.

4) Issue probe-scsi-all (or boot if the device is the boot device).

   - If you get a "Fast Data Access MMU Miss" as shown in the "Problem 
     Description" section, you have reproduced the issue.

5) Setkeyswitch standby or off.

6) Setkeyswitch on.

7) Issue reset-all.

8) Issue probe-scsi-all (or boot if the boot device, but note that the device path has changed).

   - If errors persist, your issue is different than the one described in this
     document and CR.  Perhaps you have bad hardware (card, cable, etc.).
   - If this works, Workaround #3 is confirmed.

9) Setkeyswitch standby/off.

10) Disconnect the cable from scsi interface 0 and attach it to scsi interface 1 on the I/O card (assumes Cauldron-S; SCSI 0 is top port, SCSI 1 is bottom port).

11) Setkeyswitch on

12) Issue probe-scsi-all (or boot if the boot device, but note that the device path has changed).

   - If errors persist, your issue is different than the one described in this
     document and CR.  Perhaps you have bad hardware (card, cable, etc.).
   - If this works, Workaround #1 is confirmed.

13) Setkeyswitch standby/off.

14) Move the I/O card to an empty slot on the same I/O Board. Attach the array to either scsi port (or repeat this later to test each port).

15) Setkeyswitch on.

16) Issue probe-scsi-all (or boot if the boot device, but note that the device path has changed).

   - If errors persist, your issue is different than the one described in this
     document and CR.  Perhaps you have bad hardware (card, cable, etc.).
   - If this works, Workaround #2 is confirmed.