Bogus CPU usage

cyberguypr

Before I go down the rabbit hole chasing this, does anyone have an idea what is going on here?



We have 4 identical BL460c blades and for some reason one is reporting this weird CPU usage. We wiped it and gave it a fresh install of 5.5 but the issue remains.

Thanks in advance for any pointers.

Essendon

I am not sure what's going on, but have you done a firmware upgrade of the blades/matrix at all? Possible firmware mismatch? Try plonking it into a spare bay if there's one?

jibbajabba

We used 460Cs as well and I never seen that before .. That is so random .. I presume the VM view of that host doesn't show anything out of the ordinary ?

cyberguypr

We actually updated blade and chassis firmware last week. Issue was still present then. It has the latest and greatest. Unfortunately no open bays, although I could swap a couple, but since they are offsite and need to open a ticket with the colo, that sounds like a Monday morning thing

VMs are cruising, doing nothing. Only host CPU usage alert gets triggered.

Random fact: I just noticed the VMs were set by a mysterious hand to DHCP and grabbed IPs on the APIPA range.

tomtom1

Throw the host in maintenance mode, and see what it does then?

cyberguypr

Issue still happens in maintenance mode. There was one VM on the box and I migrated it to simplify the equation. No dice.

Checked ESXTOP for giggles:

tomtom1

Your %USED and %WAIT on the hostd is very high! Try restarting some core services (put the host in maintenance mode first, so no active VM's are on the host).

~ # services.sh restart

cyberguypr

Been there done that with no effect whatsoever. Just to recap, there are no VMs on this host.

LeifAlire

Maybe the vendor sent you a beta cpu that will be released at a later date

netsysllc

Sounds like a bad CPU possibly, I had one recently I thought was a bad CPU but turned out to be fixed with a BIOS update Geek Babble | Unexplained High CPU usage on Hyper-V host and Guests

cyberguypr

Quick update: Working with HP but all testing comes back clean. Nothing points to hardware issues.

Also tried running Ubuntu live. CPU seems to be cruising there.

jibbajabba

Long shot but tried to reset the bios ?

cyberguypr

Yeah, done that two times, before and after the firmware update.

TheProf

Have you tried removing the shared storage temporarily? I had an issue once where for some odd reason, adding a LUN to the ESXi host would crash host due to high CPU usage, couldn't figure what it was, tried different storage platforms too, same result. Ended up replacing the blade all together.