CentOS 7 on a SuperMicro Server -- 'Zombie System'

[SeaweedBrain]

Hello!

I've been running into a strange problem on several SuperMicro systems and I'm trying to rule out if it's something we've done with the OS.
The system is a SuperMicro 4028GR-TR running CentOS 7. We've also got our own automated test software that runs pretty deep into the system, and uses IPMI for automatic power cycling.

The issue is:
When I try to power down the system, through GUI or command line, the system will power off regularly for about 5 seconds, then power itself back on. I've been working with SuperMicro as well to troubleshoot the system side of things.

I've been calling it 'Zombie System Behavior.' I'd appreciate any insight for things to check around in the OS.

Thanks!

[lightman47]

I have an HP Elitebook 8460p laptop that does exactly this when I try to shut it down via software. The problem used to be intermittent but now is 100% reproducible. But this same laptop now has developed other issues (battery charging/lack thereof) that all point to 'firmware' issues.

Is all your firmware up to date (BIOS), etc.?

My personal suspicion is power management code.

[ron7000]

I have a bunch of 2u X10 systems socket 2011, and a few 4029 systems like your current.
I've run/tested rhel 7.6 and centos 7.6 successfully without running into a problem like u described, so I would agree it's somewhere on supermicro's end with IPMI and power management. I can do "init 0" and "init 6" from within linux all day long with zero problems. And I don't mess with IPMI much on the older 2u systems I have but on the newer 4029 the IPMI was not 100% problem free so...

and i did a supermicro bios & ipmi update on my 4029's and did not correct the problem i was having (which was nothing major because i forget what it even was)