Ubuntu 18.04 LTS (Bionic Beaver) generates the following errors when booting in VMware Fusion 11.1.0 (13668589) on macOS 10.12.6 (16G2016) Sierra on iMac 27‑inch Retina 5K 2017:
[ 1.546521] piix4_smbus 0000:00:07:3: SMBus Host Controller not enabled!
[ 2.101002] sd 2:0:0:0: [sda] Assuming drive cache: write through
WARNING: Failed to connect to lvmetad. Falling back to internal scanning.
Gave up waiting for suspend/resume device
/dev/mapper/ubuntu--vg-root: clean 252473/1872304 files, 2142129/7476224 blocks"
Eventually, it shows the Ubuntu desktop.
How to fix it? Thanks.
For the SMBus Host Controller error, add it to the blacklist:
As for the lvmetad error, afaik it is a kernel bug that also happens on real hardware.
Thanks. I wonder if doing as you indicate would have negative effects or somehow compromise the system in the future. I mean, I know almost nothing about Ubuntu and do not want to break something now or in the future. Do you think that these issues will be factory-fixed by new versions of Ubuntu or VMware in the future? For me that would be the best and safest approach. Thanks again.
I doubt VMware will create a virtual SMBus Host controller, as it makes little sense to make a virtual system management bus controller apart from getting rid of that error.
As for the lvmetad... it's been an warning/error since at least 3 years according the debian bugtracker.. (it isn't for vmware to fix)
Down here I see those warnings at boot too and I chose to ignore them. It's more informational than anything else.
The SMBus host error has been there as long as I can remember.
See here an example of the problem as described for Ubuntu 8.x
The lvmetad error however I think I have only seen on a VM of mine which has an encrypted lvm partition.
I don't always use lvm, so don't know how common it is to see that warning.
If the lvm partition works though then I doubt it is that bad.
AFAIK, the lvmetad is a daemon that caches information about LVM layout for the lvm command line tools and if the daemon isn't running then there's no cache and it will have to scan for finding the layout.
See here for what it does:
and the man page:
The encryption part is just where I was seeing it. As mentioned before I rarely use lvm containers. On a VM it isn't offering me much beyond just an extra layer.
VM disks are easy to grow in contrary to physical disks where LVM can be pretty useful. The reason I opted for lvm on that machine was because I needed to encrypt customer data.
Doing that via an encrypted lvm container is the better solution imo.
So yes, just using lvm is enough to be able to trigger this warning.
Oh forgot to say that using lvm is a default nowadays on distributions such as ubuntu/debian/RHEL, so it isn't strange to have a disk with lvm containers.
Well, it won't simplify the troubleshooting...
This link gives a bit background info:
If you do a V2P and your sensors won't work then the kernel module not showing up in list of loaded kernel modules might be a good hint.
As I mentioned earlier, I just ignore the error.
I know what it is about and it is "fine", but I can understand that people rather see a boot without errors, for a variety of reasons.