Closed geissonator closed 3 years ago
Some further digging found that my witherspoon system had a weird /etc/systemd/network/00-bmc-eth0.network file. It was configured for a static IP, but also had "DHCP=ipv6" in it. I changed that to "DHCP=no" and all is good now. On my rainier, it looks like we have a legit issue with bringing up DHCP on eth1.
I've seen this intermittently for a while so I finally dug into it a bit. The symptom is that the systemd-networkd-wait-online service will fail.
In most cases, it seems like this service runs early enough that it passes (i.e. the network interface(s) get into a passable state. But in most cases, later in the boot to BMC Ready, they go back into states that would cause this server to fail. If the service happens to run a bit later in the boot, it will fail due to this.
Here's a witherspoon system running master (2.10.0-dev-1257-g90f181d3e). System is at Ready state and if I run the service directly, it fails:
The issue appears to be that eth0 is stuck in "configuring"
Here's what I hope are the relevant journal entries for this:
I need to dig into the systemd service a bit more, I'm wondering if this is what's getting it stuck?