Dell iDRAC 9: Alerting, RAID Configuration, and Hardware Monitoring

Parts 1 and 2 of this series established iDRAC access and hardened the management plane. Now we move into the operational configuration that makes a server production-ready: alerting and storage. A server that can’t notify your team of a hardware fault — or whose RAID array isn’t properly configured before OS installation — is a liability waiting to surface at the worst possible time.

This part covers configuring email, SNMP v3, and syslog alert destinations with verified test delivery; building RAID virtual disks and assigning hot spares via RACADM; and establishing a hardware monitoring baseline across temperature, power, and ECC memory sensors.

Step 6: Alerting and Notifications

📋 Runbook Callout
Setting up? Alerting is a go/no-go gate for production readiness. Do not hand off a server without verified alert delivery.
Firefighting? If you’re not receiving alerts for a known event, use the built-in test alert to isolate whether the problem is iDRAC config, network routing, or the receiving platform.

6a. Configure Email (SMTP) Alerts

In the web UI: Configuration > Alerts > Alert Settings > Email.

racadm set iDRAC.RemoteHosts.SMTPServerIPAddress smtp.corp.example.com
racadm set iDRAC.RemoteHosts.SMTPPort 587
racadm set iDRAC.RemoteHosts.SMTPAuthentication 1
racadm set iDRAC.RemoteHosts.SMTPUserName alerts@corp.example.com
racadm set iDRAC.RemoteHosts.SMTPPassword 'smtp-password'

# Configure alert destination 1
racadm set iDRAC.EmailAlert.1.Address oncall@corp.example.com
racadm set iDRAC.EmailAlert.1.Enable 1

6b. Configure SNMP Traps

SNMP v3 is preferred over v2c — it adds authentication and encryption, at the cost of additional configuration complexity (per-user credentials on both iDRAC and your trap receiver). SNMP v2c is simpler to set up but transmits community strings in plaintext, making it unsuitable for untrusted network segments. SNMP v3 requires Enterprise or Datacenter license.

# SNMP v3 trap destination
racadm set iDRAC.SNMP.Alert.1.DestIPv4Addr 10.10.0.50
racadm set iDRAC.SNMP.Alert.1.State Enabled
racadm set iDRAC.SNMP.Alert.1.SNMPv3Username snmpadmin

# SNMP v2c fallback (change community string from default 'public')
racadm set iDRAC.SNMP.AgentCommunity MonitoringStr0ng

6c. Syslog Forwarding

Requires Enterprise or Datacenter license.

racadm set iDRAC.SysLog.Server1 10.10.0.60
racadm set iDRAC.SysLog.Port 514
racadm set iDRAC.SysLog.Enable 1

For Splunk or Graylog, point this at your configured syslog input listener and verify that the receiver is accepting messages from the iDRAC source address.

6d. Select Event Categories

Enable at minimum these categories for production servers:

Category	Recommended severity	Rationale
System Health	Critical + Warning	CPU, memory, voltage faults
Storage	Critical + Warning	RAID failures, predictive disk failures
Thermal	Critical	Thermal shutdown prevention
Power	Critical	PSU failures
Audit	Informational	Login events, config changes

In the web UI: Configuration > Alerts > Alert Settings. Each category has independent severity checkboxes per destination type.

6e. Test Alert Delivery

racadm testalert -l 1
# Sends a test alert to all configured destinations

Verify receipt in your email inbox, SNMP trap receiver, and syslog platform before declaring alerting operational.

Step 7: RAID Configuration and Storage Management

📋 Runbook Callout
Setting up? Build your storage configuration here before OS installation.
Firefighting? For an active RAID failure, jump directly to RAID Troubleshooting.

7a. Navigate to Storage

iDRAC web UI > Storage. The left panel shows your PERC controller (e.g., PERC H730P Mini, PERC H745), any attached enclosures, virtual disks, and physical disks.

Identify your controller:

racadm storage get controllers
# Output shows controller FQDD, e.g.: RAID.Integrated.1-1

7b. Create a Virtual Disk

racadm storage createvd:RAID.Integrated.1-1 \
  -rl r10 \
  -pdkey:Disk.Bay.0:Enclosure.Internal.0-1:RAID.Integrated.1-1,Disk.Bay.1:Enclosure.Internal.0-1:RAID.Integrated.1-1,Disk.Bay.2:Enclosure.Internal.0-1:RAID.Integrated.1-1,Disk.Bay.3:Enclosure.Internal.0-1:RAID.Integrated.1-1 \
  -sz 0 \
  -ss 64k \
  -wp wb \
  -rp ra

Flags explained:

• -rl r10 — RAID level (r0, r1, r5, r6, r10, r50, r60). Trade-off: RAID 10 offers the best write performance and fast rebuild times but uses 50% of raw capacity. RAID 5/6 are more space-efficient but have higher write overhead (read-modify-write penalty) and significantly longer rebuild times — increasing the window of exposure to a second disk failure. Use RAID 10 for databases and high-write workloads; RAID 5/6 for archival or read-heavy workloads where capacity efficiency matters more.
• -sz 0 — use all available space
• -ss 64k — 64KB strip size. Trade-off: Smaller strip sizes (64KB) suit random I/O workloads (databases, VMs); larger strip sizes (256KB+) improve throughput for large sequential workloads (backups, media). Changing strip size after creation requires destroying and rebuilding the virtual disk.
• -wp wb — write-back cache. Trade-off: Write-back (wb) improves write performance by acknowledging writes once they hit the controller cache, but risks data loss on unexpected power loss if no battery-backed cache unit (BBCU) or capacitor-backed cache is present. Use write-through (wt) if your controller lacks a BBCU — it is slower but safe.
• -rp ra — read-ahead cache policy

Apply the job:

racadm jobqueue create RAID.Integrated.1-1 -s TIME_NOW -e TIME_NA

7c. Assign a Hot Spare

# Global hot spare (protects any virtual disk on this controller)
racadm storage assignhotspare:Disk.Bay.4:Enclosure.Internal.0-1:RAID.Integrated.1-1 \
  -type global

Trade-off: A global hot spare automatically rebuilds any failed virtual disk on the controller, making it the most flexible option. A dedicated hot spare (-type dedicated) is reserved for a specific virtual disk, which is useful when you have mixed disk sizes or want to guarantee a spare is available for a critical array. The downside of dedicated spares is that they cannot protect other arrays if the designated array is healthy — a global spare is generally preferred unless disk heterogeneity requires otherwise.

7d. Monitor Rebuild Progress

racadm storage get vdisks -o -p RemainingRedundancy,Progress

In the web UI: Storage > Virtual Disks — the degraded virtual disk shows a progress percentage and estimated completion time.

Step 8: Hardware Monitoring — Sensors, Thermal, and Power

📋 Runbook Callout
Setting up? Check baseline sensor readings after initial boot to confirm the system is healthy before OS installation.
Firefighting? Cross-reference sensor timestamps with the Lifecycle Controller log to pinpoint when a fault began.

8a. Dashboard Overview

The iDRAC Dashboard (iDRAC web UI > Dashboard) shows a color-coded health summary: green (OK), amber (Warning), red (Critical). Drill into any amber or red indicator immediately.

8b. Temperature and Fan Sensors

iDRAC web UI > System > Hardware > Fans (or Sensors > Temperature).

Via RACADM:

racadm getsensorinfo
# Lists all sensors: temperature, fan, voltage, current, with current values and thresholds

8c. Configure Thermal Profile

# Options: DefaultFanSpeedOffset, MaximumPerformance, MinimumPower, Custom
racadm set System.ThermalSettings.ThermalProfile MaximumPerformance

Trade-off: MaximumPerformance runs fans at higher speeds to keep thermals low, which increases acoustic noise and power consumption but provides the most thermal headroom — recommended for hot or dense environments. MinimumPower reduces fan speeds aggressively to save power and reduce noise, but narrows the thermal margin; only use this in well-cooled, low-ambient-temperature environments. DefaultFanSpeedOffset is a balanced middle ground suitable for most standard datacenter deployments.

8d. Power Capping

Enforce a per-server power cap to stay within rack PDU limits:

racadm set System.Power.Cap.Enable 1
racadm set System.Power.Cap.Watts 400
# Replace 400 with your target cap in watts

Trade-off: Power capping protects your PDU budget and prevents tripped breakers in dense racks, but the server will throttle CPU and memory performance once the cap is reached under heavy load. Set the cap conservatively enough to protect the PDU, but high enough to avoid throttling during normal peak workloads. Monitor actual power draw under load before finalizing the cap value.

8e. ECC Memory Error Monitoring

iDRAC web UI > System > Hardware > Memory — each DIMM slot shows correctable and uncorrectable ECC error counts.

racadm getsysinfo -d
# Includes memory module details and error counts

A DIMM accumulating correctable ECC errors is degrading — schedule replacement. A DIMM with uncorrectable ECC errors requires immediate action; the OS is likely already experiencing instability.

Step 9: Troubleshooting Common Issues

Alerts Not Being Received

• Run the test alert first: racadm testalert -l 1. If the test alert is not received, the problem is in iDRAC config or network routing — not in the event pipeline.
• SMTP: Confirm the SMTP server address, port, and credentials are correct. Check whether your mail relay requires TLS on port 587 vs. plain on port 25. Verify the iDRAC management NIC can reach the SMTP server (racadm ping smtp.corp.example.com).
• SNMP: Confirm the trap receiver IP is reachable from the iDRAC management network. For SNMP v3, verify the username and authentication/privacy credentials match exactly on both iDRAC and the receiver. Check that the receiver is listening on UDP 162.
• Syslog: Confirm the syslog listener IP and port are correct and that UDP 514 is not blocked by a firewall between the iDRAC management VLAN and the log collector.
• License check: SNMP v3 and syslog forwarding require Enterprise or Datacenter license. Verify under iDRAC web UI > iDRAC Settings > Overview > License.

RAID Virtual Disk Creation Fails

• Check disk state: All physical disks used in the virtual disk must be in Ready state. Run racadm storage get pdisks -o -p State to confirm. Disks in Non-RAID, Foreign, or Failed state cannot be used until converted or cleared.
• Foreign configuration: If disks show Foreign state (previously part of another array), clear the foreign config: racadm storage clearforeignconfig:RAID.Integrated.1-1 — this is destructive; confirm no data needs to be recovered first.
• Minimum disk count: Verify you have enough disks for the chosen RAID level (e.g., RAID 10 requires a minimum of 4 disks; RAID 5 requires a minimum of 3).
• Job queue conflicts: If a previous job is pending or failed, it may block new jobs. Check with racadm jobqueue view and clear stale jobs with racadm jobqueue delete -i JID_<id>.

Sensor Readings Look Anomalous

• Inlet temperature spikes: Check for blocked airflow, missing blanking panels, or a failed fan. Cross-reference fan RPM readings — a failed fan will show 0 RPM or a critical alert.
• Voltage sensor warnings: Can indicate a failing PSU or a power event. Check the Lifecycle Controller log (racadm lclog view) for correlated PSU events.
• All sensors show N/A or unavailable: iDRAC may not have completed hardware inventory after a reboot. Wait 2–3 minutes and refresh. If the issue persists, check iDRAC firmware version and consider a graceful iDRAC reset: racadm racreset.

ECC Memory Errors Increasing

• Correctable errors on a single DIMM: Isolate the slot using the web UI (System > Hardware > Memory). Schedule a maintenance window to reseat or replace the DIMM. Monitor error rate — a sudden increase indicates accelerating degradation.
• Uncorrectable errors: Treat as a P1 incident. The OS may be experiencing crashes or silent data corruption. Replace the DIMM immediately. If the server is in production, initiate a failover before replacement.
• Errors spread across multiple DIMMs: May indicate a CPU memory controller issue rather than individual DIMM failures. Escalate to Dell support with a full SupportAssist log bundle.

With alerting verified and storage configured, the server’s operational foundation is solid. Part 4 covers the remaining setup steps: running a firmware update pass through the Lifecycle Controller, exporting a System Configuration Profile for rapid cloning to additional nodes, and using the virtual console and virtual media for remote OS installation and incident response.

Continue to Part 4: Dell iDRAC 9: Lifecycle Controller, Firmware Updates, Virtual Console, and Automation →