Partition Bad Disk 3.3.2 Serial Number 13
starting in oracle exadata system software release 188.8.131.52, if there is a disk failure, then oracle exadata system software sends an alert stating that the disk can be replaced, and, after all data has been rebalanced out from that disk, turns on the blue ok to remove led for the hard disk with predictive failure. in oracle exadata system software releases earlier than 11.0, the amber fault-service required led was turned on for a hard disk with predictive failure, but not the blue led. in these cases, it is necessary to manually check if all data has been rebalanced out from the disk before proceeding with disk replacement.
for example, a hard disk status equal to
failed (the status for failed hard disks was
critical in earlier releases), or
warning - predictive failure is probably having problems and needs to be replaced. the disk firmware maintains the error counters, and marks a drive with
predictive failure when internal thresholds are exceeded. the drive, not the cell software, determines if it needs replacement.
if the replica store is online and connected to the active data grid, then at the time of the flash failure, oracle exadata system software determines the replica copy status. if the data in the replica store is being replicated from the online standby, then oracle exadata system software selects the replica copy. the grid disk status is
active -- replicating working. if the replica store is not available or the replica store data is not in the replica store, then oracle exadata system software selects the last page read from the target mirror.
Use this procedure to troubleshoot the problem if you have not already run the vendor-specific diagnostic utility to check the flash disk. Verify whether the alert is still displayed. Check if you have sufficient free space on the grid and cell disks to expand the remaining clusters in the existing clusters on the grid and cell disks. If you have insufficient free space on the cell disks and need to expand the clusters in the existing cell clusters on the cell disks, then resize the existing cell clusters on the cell disks. Resize the existing cell clusters on the cell disks to the desired size, and then repeat steps 1 through 3 in Partition Bad Disk 3.3.2 Serial Number 13. Starting with Oracle Exadata Database Machine release 184.108.40.206.0, if there is a disk failure, then Oracle Exadata Database Machine sends an alert that the hard disk can be replaced, and, after all data has been rebalanced out from that disk, turns on the OK to Remove LED for the hard disk with predictive failure. In Oracle Exadata Database Machine releases earlier than 220.127.116.11.0, the amber Fault-Service Required LED was turned on for a hard disk with predictive failure, but not the blue LED. In these cases, it is necessary to manually check if all data has been rebalanced out from the disk before proceeding with disk replacement. If the predictive failure has been turned into a fault, then the cell software does not attempt to fix the disk, and, instead, displays an alert indicating a low health condition. See the Alert section of the /opt/oracle/dasw/pxfs/XE/cellsrv/cellsrv_err.log file for further details. 5ec8ef588b