{"id":11240,"date":"2019-04-01T06:57:38","date_gmt":"2019-04-01T06:57:38","guid":{"rendered":"https:\/\/dbtut.com\/?p=11240"},"modified":"2019-04-01T06:57:38","modified_gmt":"2019-04-01T06:57:38","slug":"replacing-a-damaged-hard-disk-on-exadata","status":"publish","type":"post","link":"https:\/\/dbtut.com\/index.php\/2019\/04\/01\/replacing-a-damaged-hard-disk-on-exadata\/","title":{"rendered":"Replacing a Damaged Hard Disk on Exadata"},"content":{"rendered":"
You can perform the hard disk replacement on Exadata cellnodes by performing the following checks and steps.<\/p>\n
Hard Disks on the cell node are automatically dropped at the time of damage, and the disks of the asm disk groups will also be deleted. After the disk drop operation and after the disk changes, rebalance process will automatically start in ASM disk groups.<\/p>\n
Find the Damaged Disk<\/h2>\n
CellCLI> LIST PHYSICALDISK WHERE diskType=HardDisk AND status LIKE \".*failure.*\" DETAIL\r\nname: 20:11\r\ndeviceId: 8\r\ndiskType: HardDisk\r\nenclosureDeviceId: 20\r\nerrMediaCount: 980\r\nerrOtherCount: 0\r\nforeignState: false\r\nluns: 0_11\r\nmakeModel: \"SEAGATE ST360057SSUN600G\"\r\nphysicalFirmware: 0A25\r\nphysicalInsertTime: 2011-11-29T13:40:05+02:00\r\nphysicalInterface: sas\r\nphysicalSerial: E1EY5Z\r\nphysicalSize: 558.9109999993816G\r\nslotNumber: 11\r\nstatus: predictive failure<\/pre>\n
You can also detect the damaged disk from the alert history output.<\/p>\n
CellCLI> list alerthistory\r\n1_1 2017-09-09T15:41:00+03:00 critical \"Hard disk status changed to predictive failure. Status : PREDICTIVE FAILURE Manufacturer : SEAGATE Model Number : ST360057SSUN600G Size : 600GB Serial Number : E1EY5Z Firmware : 0A25 Slot Number : 11 Cell Disk : CD_11_orclcel08 Grid Disk : RECO_ORCL_CD_11_orclcel08, DBFS_DG_CD_11_orclcel08, DATA_ORCL_CD_11_orclcel08\"<\/pre>\n
Check the Rebalance Operation in ASM Disk Groups<\/h2>\n
Do not change the disk if there is a running rebalance operation. Wait for the process to finish.<\/p>\n
[root@orcldb01 ~]# su - oracle\r\n[oracle@orcldb01 ~]$ . oraenv\r\nORACLE_SID = [oracle] ? +ASM1\r\nThe Oracle base has been set to \/u01\/app\/oracle\r\n[oracle@orcldb01 ~]$ sqlplus \/ as sysasm\r\n\r\nSQL*Plus: Release 11.2.0.3.0 Production on Mon Sep 11 15:46:58 2017\r\n\r\nCopyright (c) 1982, 2011, Oracle. All rights reserved.\r\n\r\nConnected to:\r\nOracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production\r\nWith the Real Application Clusters and Automatic Storage Management options\r\n\r\nSQL> select * from gv$asm_operation;\r\n\r\nno rows selected\r\n\r\nSQL> exit\r\nDisconnected from Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production\r\nWith the Real Application Clusters and Automatic Storage Management options<\/pre>\n
Check the Status of the Disk to be Replaced<\/h2>\n
CellCLI> list celldisk where lun=0_11\r\nCD_11_orclcel08 proactive failure\r\n\r\nCellCLI> list griddisk where celldisk=CD_11_orclcel08 attributes name,size,status,asmmodestatus\r\nDATA_ORCL_CD_11_orclcel08 423G proactive failure DROPPED\r\nDBFS_DG_CD_11_orclcel08 29.125G proactive failure DROPPED\r\nRECO_ORCL_CD_11_orclcel08 105.6875G proactive failure DROPPED<\/pre>\n
Make sure that the Asm Disks on the Damaged Disk are Removed from the Disk Groups<\/h2>\n
With the following query, you can be sure that there are no asm disks associated with the damaged disk. Because the values \u200b\u200bin the name column in the v$asm_disk view are in the upper case, use upper case when querying.<\/p>\n
SQL> select * from v$asm_disk where name like '%CD_11_ORCLCEL08%'<\/pre>\n
Although it is not necessary, it is useful to remove the disk completely from the system with the following command. In some versions, this command may cause a syntax error. If values from the above query returns, the following command must be executed.<\/p>\n
CellCLI> alter physicaldisk 20:11 drop for replacement\r\n\r\nPhysical disk 20:3 was dropped for replacement.\r\n\r\nCellCLI>\r\n\r\nCellCLI> list physicaldisk 20:11 detail\r\nname: 20:11\r\ndeviceId: 8\r\ndiskType: HardDisk\r\nenclosureDeviceId: 20\r\nerrMediaCount: 980\r\nerrOtherCount: 0\r\nforeignState: false\r\nluns: 0_11\r\nmakeModel: \"SEAGATE ST360057SSUN600G\"\r\nphysicalFirmware: 0A25\r\nphysicalInsertTime: 2011-11-29T13:40:05+02:00\r\nphysicalInterface: sas\r\nphysicalSerial: E1EY5Z\r\nphysicalSize: 558.9109999993816G\r\nslotNumber: 11\r\nstatus: warning - predictive failure - dropped for replacement<\/pre>\n
Change the Damaged Disk<\/h2>\n
You can change the disk at this stage. When ejecting the disc, click the button and pull the disk a little and wait for the lights to turn off completely. Then remove the disc completely from the slot.<\/p>\n
After the change, you should check the following to see if everything is normal.<\/p>\n