Node services are hung and all drives become available after a Correctable ECC error

Applies to

NetApp H700S

Issue

• System Event Log (SEL) at BMC/IPMI shows Correctable ECC alert at the time of event
  • <Event ID> | <DATE> | <TIME> | Memory | Correctable ECC (@DIMM<ID>(CPU<#>)) | Asserted
• NetApp SolidFire ActiveIQ and cluster UI shows that all block and metadata services of the node are unresponsive shortly after the SEL log entry:
  • unresponsiveService - A block service is not responding. (This alert will be shown for each block drive)
  • unresponsiveService - A metadata service is not responding.
• Unhealthy service alerts are shown  shortly after the unresponsive services alerts:
  • blockServiceUnhealthy - A block service is unhealthy and SolidFire is attempting to migrate data away from it.
  • sliceServiceUnhealthy - A metadata service is unhealthy and SolidFire is attempting to migrate data away from it.
• All drives of the affected node becomes available after a while (NetApp SolidFire ActiveIQ and cluster UI):
  • driveAvailable - Node ID X has Y available drive(s).