EXADATA Flash Disk Replacement - No downtime for databases.

Below technical steps will be followed to replace the flash disk in exadata server. This activity require a restart of the cell server but there will be no downtime for databases.


Log onto cell server as a root-  
                                                                                
#cellcli -e list griddisk attributes name,asmmodestatus,asmdeactivationoutcome

Above command must say asmmodestatus= 'ONLINE' and asmdeactivationoutcome= 'Yes' for all griddisks

Example :

DATA_CD_00_cell01 ONLINE Yes

DATA_CD_01_cell01 ONLINE Yes

DATA_CD_02_cell01 ONLINE Yes

RECO_CD_00_cell01 ONLINE Yes

RECO_CD_01_cell01 ONLINE Yes

RECO_CD_02_cell01 ONLINE Yes

#cellcli -e alter griddisk all inactive

#cellcli -e list griddisk attributes name,asmmodestatus,asmdeactivationoutcome
Above command should say 'OFFLINE ' for all grid disks

--> during the transition:

DATA_CD_00_cell01 SYNCING No

DATA_CD_01_cell01 ONLINE   No

RECO_CD_00_cell01 OFFLINE  Yes

RECO_CD_01_cell01 ONLINE  Yes

--> Finally, you might see something like

DATA_CD_00_cell01 OFFLINE Yes

DATA_CD_01_cell01 OFFLINE Yes

RECO_CD_00_cell01 OFFLINE Yes

RECO_CD_01_cell01 OFFLINE Yes

To see whether the disks were actually inactivated, a commonly used command is:

#cellcli -e list griddisk

#cellcli -e "list griddisk where active=FALSE"

Above command should say 'inactive'

DATA_CD_00_cell01

DATA_CD_01_cell01

RECO_CD_00_cell01

RECO_CD_01_cell01

Power off the cell server
#shutdown -h now
Above command will shutdown the cell server


REPLACE FLASH DISK - Hardware replacement
Oracle Engineer slides the cell out from the rack
Oracle engineer removes the Flash card and inserts the new one


Cell server is powered back on
At ILOM fire the below command
start /SYS
Above commd will start the cell server


Disk sync is carried out


#cellcli -e alter griddisk all active
#cellcli -e list griddisk
Above command should show 'active'


#cellcli -e list griddisk attributes name, asmmodestatus
Above command will say OFFLINE then SYNCING then ONLINE


Full resilience is restored  and there is no interruption to service during the replacement.