Skip to main content
Hardware

Fujitsu ServerView Health (SNMP)

Monitor the hardware health of your Fujitsu PRIMERGY servers over SNMP through the ServerView agents, from temperatures to RAID.

The Fujitsu ServerView Health (SNMP) check monitors the hardware health of Fujitsu PRIMERGY servers over SNMP — overall system status, CPU and mainboard, temperatures, fans, power supplies, memory, and RAID. It gives you early warning when a component starts to fail. The check runs on an agent device on Windows, Linux, or macOS (each in x64 and Arm64 builds) and queries the server over the network, every 15 minutes by default.

Prerequisites

  • You need the Monitoring Check Management permission. Permissions are assigned under Administration → Groups.
  • An octoja agent on a device with network access to the Fujitsu server.
  • Fujitsu ServerView Agents installed and running on the server's operating system. This package provides the ServerControl 2 (SC2) SNMP agent and the ServerView status rollup (SVS); the check reads model and serial number, overall status and subsystems, temperatures, fans, power supplies, and memory from them. ServerView Update Agent is not required for this check.
  • For the list of physical disks, additionally ServerView RAID Manager or the SNMP agent from your RAID controller's software package. According to the Fujitsu hint on SNMP agents for RAID controllers, these agents are not included in the ServerView Agents package; they are installed automatically only when Windows is initially installed through ServerView Installation Manager. Without the RAID agent the check still runs but shows no physical disks.
  • SNMP enabled on the operating system and reachable on the configured SNMP port — UDP 161 by default — with a community string (v1/v2c) or USM user and keys (v3).
  • Point the check at the server's operating system address — not the iRMC management controller.

Set up SNMP on a Windows server

The check talks to the SNMP service of the server's operating system; the ServerView agents plug into that service. On a Windows server you therefore need, regardless of the Windows version:

  • The Windows SNMP service as an installed feature. It is not preinstalled; you add it through the optional Windows features.
  • A read-only community for SNMP v1 or v2c. This is exactly the string you enter in the check's SNMP Community field; the check needs no write access.
  • The IP address of the device running the octoja agent that executes the check as a permitted query host in the SNMP service's security settings. By default Windows accepts requests from the local machine only; without this entry the check gets no answer.
  • An inbound allow rule for UDP 161 in the server's Windows firewall and on any firewall in between, each from that device.

The exact dialogs differ between Windows versions. The installation order of the ServerView packages and how they interact with the SNMP service are documented in the Fujitsu manual for the respective package on support.ts.fujitsu.com. Fujitsu starts and stops the ServerView Agents together with the SNMP service: after changing the service, restart both.

Steps

You can assign the check to a network device as well as to an agent device. On an agent device, add it to the device that reaches the Fujitsu server over the network.

  1. Go to Devices and open the device that has network access to the Fujitsu server.
  2. Click the Checks tab → Add Check → Fujitsu ServerView Health (SNMP).
  3. Enter the server's hostname or IP address in the Server host / IP (OS, not iRMC) field. Use the address of the server's operating system, where the ServerView agents run.
  4. Choose the SNMP version the server is configured for. For v1 or v2c, enter the SNMP Community. For v3, enter the user and select the authentication and privacy protocols and passwords.
  5. (Optional) Set Extra temp warning cap (°C, 0 = off) in °C to add a single warning threshold on top of each sensor's own limits. Leave it at 0 to rely only on the thresholds the hardware reports.
  6. Click Add Check.

After the first check interval, the result appears in the device's Checks tab.

Settings

The table lists every field in the order it appears in the check editor.

FieldTypeRequiredDefaultDescription
Server host / IP (OS, not iRMC)StringYes—Hostname or IP of the server's operating system, where the ServerView agents run (not the iRMC)
SNMP versionSelect (v1 / v2c / v3)Yesv2cSNMP protocol version used to query the server
SNMP portNumberNo161UDP port the server answers SNMP on
SNMP CommunityStringNopublicSNMP v1/v2c community string (shown when version is v1 or v2c)
SNMPv3 userStringNo—USM user name (shown when version is v3)
Authentication protocolSelect (None / SHA-1 / SHA-256)NoNoneMessage authentication algorithm (v3 only)
Authentication passwordPasswordNo—Shown when an authentication protocol is selected
Privacy protocolSelect (None / AES-128 / AES-192 / AES-256)NoNonePayload encryption algorithm; requires authentication (v3 only)
Privacy passwordPasswordNo—Shown when a privacy protocol is selected
Extra temp warning cap (°C, 0 = off)NumberNo0Adds one global warning threshold on top of each sensor's own limits. 0 disables it

Temperatures use each sensor's own warning and critical limits as reported by the hardware. The extra temp warning cap is an additional, global safeguard and applies only to sensors that have not already reached their own warning or critical limit, so one sensor never counts twice.

Results

ConditionSeverity
Server unreachable over SNMP, or no overall status returnedCritical
Overall or subsystem status reports error or failedCritical
A sensor temperature reaches its critical thresholdCritical
Overall or subsystem status reports degradedWarning
Overall status is unknown — ServerView returned a value other than ok, degraded, error or failedWarning
A sensor temperature reaches a warning threshold, or the extra temp warning capWarning
Host left blankWarning
All subsystems report OKOK

The overall result is the worst severity across every subsystem and sensor. An individual subsystem with ServerView status unknown remains visible in the result but does not increase the overall severity. An unknown overall status still produces a warning. Inspect individual components rather than treating OK as confirmation for every subsystem.

When the check reports Unreachable

If the check shows Critical with the health Unreachable, the server either did not answer over SNMP at all or returned no ServerView overall status. Check in this order and move on only once the previous point holds:

  1. SNMP service running. The operating system's SNMP service is running on the server and listening on the port entered in the check, UDP 161 by default.
  2. Community and query host. The community in the check matches the one configured on the service, and the IP address of the device running the octoja agent is listed as a permitted host. A wrong community or a missing host looks like a timeout from the outside, because the service simply does not answer.
  3. Firewall. UDP 161 is allowed inbound from that device in the server's Windows firewall and on any firewall in between. To test, query the standard system OID 1.3.6.1.2.1.1 from the device running the octoja agent with an SNMP tool of your choice and the same credentials. If it answers, service, credentials, and firewall are fine.
  4. ServerView service running. If the system OID answers but the check stays on Unreachable, the ServerView agent returns no overall status. Verify that the ServerView Server Control service is running on the server. The OID 1.3.6.1.4.1.231.2.10.2.11.2.1.0 must return a value between 1 and 5.
  5. Only now the package question. If the ServerView OID stays empty although the service is running, the ServerView Agents are missing or incomplete; verify the installation against the Fujitsu manual for the package. If only the physical disks are missing while everything else is correct, the RAID SNMP agent (ServerView RAID Manager or the controller's software package) is missing; a query below 1.3.6.1.4.1.231.2.49 then returns nothing.

Collected metrics

The check reports:

  • Server model and serial number
  • Overall system status, and the CPU / Systemboard status read from the server's Systemboard subsystem
  • Status for every subsystem the server's ServerView rollup reports, under the names ServerView itself uses
  • Highest temperature and the sensor that reported it, plus warning and critical sensor counts
  • Fan count and lowest fan speed (RPM)
  • Power-supply count
  • Total installed memory (GB)
  • Per-disk name and capacity from ServerView RAID — the disk model stands in when the RAID agent reports no name

Tips

  • The check reads its data from the Fujitsu ServerView agents on the server, so those agents must be installed — a bare SNMP service on the operating system is not enough. Which package provides which values is listed under Prerequisites.
  • For a Hyper-V host running on the Fujitsu server, monitor the hypervisor separately with the Microsoft Hyper-V Host check.