:: Re: [DNG] udev dmi_memory_id errors
Página Principal
Delete this message
Reply to this message
Autor: Vince Mulhollon
Data:  
Para: dng
Assunto: Re: [DNG] udev dmi_memory_id errors
On Thu, Aug 29, 2024 at 11:07 AM Tom via Dng <dng@???> wrote:
> Repeated about 9 times. Does anyone know what this relates to and how to fix it?


I remember seeing that when I was running on vmware. IIRC nothing bad
happened, but that proves little about your situation. I no longer
have access to vmware, but I do have access to proxmox.

man dmidecode

and if you're happy with that manpage

sudo dmidecode -t memory

On my proxmox it looks like

Handle 0x1100, DMI type 17, 40 bytes
Memory Device
        Array Handle: 0x1000
(boring stuff deleted)
        Size: 16384 MB
        Form Factor: DIMM
        Set: None
        Locator: DIMM 0
(boring stuff deleted)
        Manufacturer: QEMU
(ha ha)


My guess is vmware is emulating a DMI reporting to your VM that you
have a DIMM with 16384 bytes (appropriate for a TRS-80 Model III) not
16384 MB

If you run this on real live physical hardware you'll see your DIMMs
report their part numbers and configured speeds and serial numbers and
all that. The days of the old 4116 dram dip that contained about 16K
of ram plus or minus some bits stuck on or off, and nothing else, are
long in the past. As you'd expect by a world-wide standard, it
doesn't work all the time. My Hynix Semiconductor DDR4 32 GB DIMMs on
my cluster hosts are running off "Configured voltage: 0.003 V" which
is impressively low LOL. All 18 DIMMs in the cluster report the same
weird voltage. The IPMI is reporting to Zabbix they are actually
running at 1.2 volts, which is much more plausible for "late 2010s
technology".

WRT it being reported 9 times I would guess that like me you're
running something like Zabbix that periodically polls the system
including the DIMMs, which is useful on a physical machine, but
probably not terribly useful on a VM.