Changing the iDRAC IP address on a Dell EMC ECS node takes one command from the node itself. The command is the easy part. What is worth knowing before you plan a maintenance window around it is that on an ECS appliance the change does not survive a reboot of that node.
The command
admin@ecs-node-04:~> sudo -i racadm setniccfg -s <IPv4Address> <netmask> <IPv4 gateway>
admin@ecs-node-04:~> sudo -i racadm setniccfg -s 192.168.20.22 255.255.255.0 192.168.20.254
Static IP configuration enabled and modified successfully
racadm runs locally on the node — you do not need the iDRAC to be reachable to change it, which is the whole point when the current address is wrong or on a network you cannot get to. (If it is reachable and you would rather drive it from your own machine, installing the iDRAC Tools racadm gets you the same commands remotely.) The new address is live within a few seconds; there is no iDRAC restart and no node restart involved.
Check it straight away:
admin@ecs-node-04:~> sudo -i racadm getniccfg
NIC Enabled = 1
IPv4 Enabled = 1
DHCP Enabled = 0
IP Address = 192.168.20.22
Subnet Mask = 255.255.255.0
Gateway = 192.168.20.254
Reassuring, and worth nothing — see below.
Note what that command shows you before you touch anything: on an ECS appliance the iDRAC and BMC addresses are deployed on the 192.168.219.x subnet, one per node, so node 4 sits at 192.168.219.104. That range is not an accident, which turns out to matter.
The catch: the iDRAC IP address comes back after a node reboot
On a plain PowerEdge server, setniccfg is persistent — it writes the iDRAC's own configuration and that is the end of it. On an ECS appliance it behaves differently: after the node is rebooted, the address is back to what it was before. The change was real, it was applied, and it is simply gone.
That makes this a perfectly good command for what most people need it for — getting into a node's iDRAC now, during an intervention — and a poor one for changing the management addressing of an appliance permanently. If you renumber a rack this way, everything works until the first node reboot, and then the addresses drift back one by one as nodes are restarted over the following months, which is a miserable thing to debug later.
Verify it the way that actually proves something
Checking with getniccfg right after the change proves nothing about persistence. The only check that does is on the other side of a restart:
sudo -i racadm getniccfg # before the reboot
# reboot the node through the normal ECS procedure
sudo -i racadm getniccfg # after it is back
If the second reading shows the old address, you have the behaviour described here and there is nothing wrong with your syntax.
Making it permanent: Dell says do not
There is documentation, and it is blunter than I expected. Dell's KB ECS: Configuring the iDRAC or BMC for out of band management with external IP addresses states that "configuring the ECS EX-Series iDRAC or BMC on IP addresses other than the defaults deployed is not supported", that "the ECS EX-Series implementation requires the iDRAC physical interface to be configured on the 192.168.219.x subnet" and that this "cannot be configured to a different physical connection or network IP subnet". Custom iDRAC addressing is listed as a future feature. The KB scopes itself to ECS Appliance Gen 3 — EX300, EX500 and EX3000.
What that KB does not say is what happens if you change the iDRAC IP address anyway, which is the gap this post fills: it reverts. The two statements are separate and worth keeping separate — Dell establishes that it is unsupported, not why it does not stick.
So open a case rather than guess. An appliance whose hardware management addresses do not match what the platform believes they are is exactly the sort of mismatch that surfaces months later during a firmware upgrade or a disk replacement.
While you are on the node, the commands that check a failed disk are worth having in the same window — racadm getsvctag is usually the next thing you want after you have reached the iDRAC.
