Best Practices for Data Protection

This article describes Promise Technology’s recommended best practice procedures for optimal data protection.

Predictive Data Migration - PDM

PDM, one of the unique technologies in the Promise Perfect Raid suite, should always be enabled. PDM proactively detects possible drive failures before they could occur and migrates data to a new healthy drive. Predictive Data Migration dramatically reduces the potential for data loss as well as the likelihood of a logical drive (array) going critical. PDM requires a global spare or unconfigured drive within the array. If PDM cannot be implemented (no spare or unconfigured drives available), there’s another proactive mechanism that can be implemented: Medium Error Threshold settings.

Medium Error Threshold

Medium Error Threshold is a configurable value within the cli and WebPAM PROe. The value is the number of bad blocks tolerated before the controller marks the drive as dead. By default, the Medium Error Threshold is disabled. The value can be set from 0 to 4294967294. The threshold only functions when the array is in an OK or Degraded state. It will not function if the array is critical or if a non-redundant RAID type (Raid 0) is used.

  • Medium Error Threshold configuration can be used in conjunction with PDM for even greater disk health monitoring.

Configuration

Configuration from the CLI:

  1. Log-in to the VTrak EClass via Telnet or the Serial console.
     
  2. When at the CLI prompt, type the following command to set the Medium Error Threshold to 64 bad blocks.
     
         administrator@cli> phydrv -a mod -s "mediumerrorthreshold=64" 
     
  3. To confirm the setting change, type the following command at the CLI prompt.
     
         administrator@cli> phydrv -v -p 1
     
  4. The resulting command output will show info for physcial drive Pd Id 1. 
     
       ----------------------------------------------------------------------
       PdId: 1 
       OperationalStatus: OK 
       Alias: 
       PhysicalCapacity: 698.64GB            ConfigurableCapacity: 698.49GB 
       UsedCapacity: 0Byte                   BlockSize: 512Bytes 
       ConfigStatus: Global Spare            Location: Encl1 Slot1 
       ModelNo: ATA     Hitachi HUA72107     VisibleTo: All Controllers 
       SerialNo: GTA200P8G247SA              FirmwareVersion: GK8OA70M 
       DriveInterface: SATA 3Gb/s            Protocol: ATA/ATAPI-7 
       WriteCacheSupport: Yes                WriteCache: Enabled 
       RLACacheSupport: Yes                  RLACache: Enabled
       SMARTFeatureSetSupport: Yes 
       SMARTSelfTestSetSupport: Yes          SMARTErrorLoggingSupport: Yes
       CmdQueuingSupport: NCQ                CmdQueuing: Enabled 
       CmdQueueDepth: 16                     MultiDMASupport: MDMA2 
       MediumErrorThreshold: 64 
       UltraDMASupport: UDMA5                DMAMode: UDMA5 
       Errors: 0                             NonRWErrors: 0 
       ReadErrors: 0                         WriteErrors: 0
     
  5. You can verify that the Medium Error Threshold setting is now set to 64 bad blocks as shown above

 

Configuration from WebPAM PROe

  1. Log-in to WebPAM PROe.
     
  2. Click on Enclosures -> Enclosure 1 -> Physical Drives and then choose the Global Settings tab.  
    You will see a screen like the example below:
     

     
  3. Type in the number of bad blocks for Medium Error Threshold. You may adjust the settings for SATA or SAS drives independantly.  For example, set the Medium Error Threshold to 64 bad blocks.
     
  4. Click the Submit button.
     
  5. When WebPAM PROe screen updates, click on the Global Settings tab again. This will display the change.
    You will see a screen like the example below:
     

 

Please note - Independant from PDM and Medium Error Threshold, a physical drive will always be marked dead if 5 timeouts are experienced within a 30 minute period.

 RelatedArticles:

“KB-10255 New Error Threshold Values for SR2.6”

http://kb.promise.com/thread/new-error-threshold-default-values-for-sr2-8