Sun System Handbook - ISO 3.4 June 2011 Internal/Partner Edition | |||
|
|
Solution Type Problem Resolution Sure Solution 1304656.1 : Hardware Alert: Could not sync the following change in disk state with Cellsrv: Lost LUNs
In this Document Created from <SR 3-3152621681>
Applies to:Exadata Database Machine X2-2 Full Rack - Version: Not ApplicableInformation in this document applies to any platform. SymptomsReceiving a Hardware Alert: CRITICAL,Could not sync the following change in disk state. ChangesNone.CauseKnown Riser issue. Sun hardware unpublished BUG: 6997590 No fix is available yet. But there is a workaround. SolutionThe impact of this problem on the operation of the machine as a whole is minimal. It is recommended to workaround this problem when it occurs as soon as it is convenient for the customer to do so. Schedule a suitable time to reboot the cell server NOTE: You can reboot the cell server wihtout any interruption to the operation of the RAC nodes. However, this introduces a risk of crashing the cluster IF another failure occurs on another cell server that happens to be the mirror partner of the server you are rebooting. If any FMOD on these cards was in use at the time of the riser disappearing, it may have alerts and status "poor performance" rather than "not present" after the reboot. To overcome this situation for any FMOD LUN in this state, force a "re-enable" on the flash cache LUNs back into the configuration using the following commands from cellcli interface: CellCLI> alter lun <flashlunname> reenable force for each lun that is marked as "poor performance" in output from the above 'list physicaldisk' command. Then drop and re-create the flash cache:
The system will be back online automatically immediately after rebooting is completed. References<BUG:6997590> - X4270 EXADATA DATA CELLS LOSING PAIRS OF AURA 1.0 FLASH CARDS AT THE SAME TIME DURING PRODUCTION.Attachments This solution has no attachment |
||||||||||||
|