VMware Fusion 10.1.3 hogs cpu - hgfs Fuse using 60% of two processors in Mint

[slgarwood]

I am running an application that accesses Mac files from Mint 19 using a shared folder -- when the app is running (pretty heavy I/O I admit) the VM is running over 60% on the two virtual cpus just to do I/O. Seems a bit excessive.

[bfan]

Hi,

What application you are running? I could have a try

[slgarwood]

The virtual system is Mint 19 running MySQL server - queries are inbound from my OSX system. Probably something you just get when doing a lot of MySQL IO on a VM under OSX ...

[Laurens73NL]

I got no solid solution for this problem, but I found a way to work-around it.

I created two shellscripts. the first one is started automaticly with Xorg (I put it in the startup list in XFCE) - this one connects to the root of the shared folder of my Mac host, see the code of this shell script below:

Code:

--------------------------------------------------------------------------------------

laurens@Mac ~/shellscripts $ cat connect_hgfs.sh

#!/bin/bash

#

# Force connection with hgfs volumes for virtual machines

#

echo mounted drive:

ls    /mnt/hgfs/

#

echo mounted directory:

ls    /mnt/hgfs/Macintosh\ HD/

#

echo testing write permissions:

touch /mnt/hgfs/Macintosh\ HD/Users/laurens/.connected_linux_machine

ls    /mnt/hgfs/Macintosh\ HD/Users/laurens/.connected_linux_machine -la

--------------------------------------------------------------------------------------

Windows clients also suffer this problem. This can be 'solved' by opening the Windows Explorer and connect with the shared drive before directly going to a deeper directory.

In the root folder an  empty text file .connected_linux_machine gets created, when the connection is succesful.

The second thing is hanging of the vmhgfs-fuse daemon, simply kill and reconnect lets the system go further.

Code:

--------------------------------------------------------------------------------------

laurens@Mac ~/shellscripts $ cat restart_hgfs.sh

#!/bin/bash

#

echo Warning: this script must be run as root or sudoer in order to achieve a system wide remount

#

killall vmhgfs-fuse

vmhgfs-fuse -o allow_other -o auto_unmount -o uid=1000 -o gid=1000 .host:/ /mnt/hgfs

echo mounted drive:

ls    /mnt/hgfs/

echo mounted directory:

ls    /mnt/hgfs/Macintosh\ HD/

echo testing write permissions:

touch /mnt/hgfs/Macintosh\ HD/Users/laurens/.connected_linux_machine

ls    /mnt/hgfs/Macintosh\ HD/Users/laurens/.connected_linux_machine -la

--------------------------------------------------------------------------------------

Of course laurens needs to be replaced by your own user login.

*Edited a few times to make it more readable

[credd]

Hi,

I too see huge CPUloads for vmhgfs-fuse in an Ubuntu 18.04 guest in VMware fusion on MAC.

It's been a consistent problem within a few hours of runtime.  The reset script above helps.

But is there a more general fix for this?  Is VMware working on vmhgfs-fuse to make sure performance is acceptable?

Thanks