I've recently taken over server administration and need advice on monitoring the hardware in a Dell server running RHEL 8. I'm comfortable with Linux and general hardware, but I'm less familiar with enterprise server management. The machine has a separate RAID array for storage, and I only have SSH access under normal circumstances.
I'm considering smartmontools for disk health, dmidecode for BIOS and memory information, and ipmitool for power-supply and fan sensors. The system already has lm_sensors for CPU temperatures, top/htop for resource usage, and ethtool for network errors. I don't need OS patch management; I mainly want reliable hardware health monitoring and alerts. Is there anything important I'm missing, particularly for the RAID controller, disks, fans, power supplies, and other Dell-specific components?
4 Answers
For a single server, Dell OpenManage Server Administrator can provide much of the hardware information through command-line tools such as `omreport` and `omconfig`. It is useful for checking the RAID controller, physical disks, virtual disks, fans, temperatures, memory, and power supplies from RHEL.
For several systems, OpenManage Enterprise is a better centralized option. It can discover the servers and collect hardware status, while iDRAC still remains the authoritative out-of-band interface for each machine.
Your proposed Linux tools are useful, but they don’t replace monitoring the RAID controller and Dell management controller. SMART data may be incomplete or unavailable for disks behind a hardware RAID controller, so make sure you check the controller’s physical-disk and virtual-disk status through OpenManage or iDRAC as well.
At minimum, alert on degraded or failed virtual disks, failed physical disks, predictive disk failures, controller faults, failed fans, temperature thresholds, power-supply failures, and memory errors. SNMP polling or traps from iDRAC can feed tools such as Nagios or LibreNMS.
You can also use the Dell MIBs with whatever SNMP monitoring system you already run. They expose useful values such as service tag, temperature and cooling probes, power-supply state, physical-disk state, and virtual-disk state. That gives you alerting without installing a large collection of agents on the RHEL host.
Since you only have SSH access, configure and test the out-of-band management network before you need it. Verify that you can reach iDRAC remotely and that alerts work; otherwise a hardware failure could leave you dependent on physical access.
The most useful Dell-specific option is iDRAC. It gives you a complete view of the physical hardware, including fans, temperatures, power supplies, memory, disks, the RAID controller, and virtual-disk health. You can monitor it remotely without needing to log into the operating system, which is especially useful when the OS is unhealthy or unavailable.
If you already have a monitoring platform, configure iDRAC to expose its data through SNMP and collect the alerts there. You can also configure iDRAC to send email notifications directly for critical events.
This is also why it’s worth checking whether the server has the iDRAC Enterprise features enabled. The basic interface may not provide every remote-management and alerting option.