Resuming a CentOS 7 guest went up from seconds to minutes. I have an APFS SSD. Is there a known issue?
One more data point. Created a 200GB disk image, HFS+, not encrypted, residing on the same disk. Suspended my VM, copied the .vmwarevm bundle into the new volume, and started it from there. Resume was under 10 seconds.
So we have the very same host, the same VM, the same disk hardware.
I think we are quite sure now that it is related to APFS+FileVault.
At least we have some workaround now.
I'm seeing slowdowns as well - I suspect there's something with APFS that's causing it. It's not just Fusion - it's any operation on a large disk file.
You are probably right. It is very visible with a 8GB suspend/resume image.
I'm seeing the same thing on High Sierra, APFS on SSD, running Fusion 8.5.8. My Windows 10 VM with 3GB of RAM takes minutes to restore.
[FIX]
The following has worked for me to resolve this issue.
1. Shutdown and power off the virtual machine that is causing the bother.
2. Go to the "Virtual Machine" menu and select the "Power On to Firmware" option.
3. Go to the "Advanced" menu in the BIOS.
4. Change "Large Disk Access Mode" to "Other".
5. Save and exit the BIOS.
6. Reboot your VM, wait for it to load and then try out the suspend/ resume process.
Please let me and others know if this also works for you.
All the best,
Scott
Thanks Scott. It fixes the issue. Indeed, it might even be faster to resume now than before the upgrade to high sierra.
I just gave this a try but unfortunately still having the same issue. Very slow resume. I think this is a VM10/APFS issue.
Glad it helped!
Unfortunately, over time it has gotten slow again. I might give linux hibernation a chance instead of vmware suspend/resume.
In my case it's slow regardless of guest OS and with or without Scott's fix. On the same VM that used to take 30 seconds to load, I have measured at just over 3 minutes and 40 seconds.
Me too.
The advice from Scott did not help in my case. Resuming a Windows 7 VM seems to take forever since the upgrade from Fusion v8.5 Pro to v10 Pro. The host machine is a MacBook Pro running macOS 10.13.1 from SSD, so we are talking APFS.
This is getting worse and worse.
Hey VMWare, when are you going to fix this issue? It's now taking over 10 minutes to start a VM that in prior releases would take 30 seconds.
+1. For a particular Windows 10 VM I use, I've found it's dramatically faster to shut down and restart every time (< 1 min to boot), rather than trying to sleep and resume (anywhere from 1 min to 10 mins to resume). Super annoying.
So I made some more tests. Resume is slow if and only if the host uses APFS with FileVault encryption. Without FileVault, resume speed is OK. Mohannad and gr2020, are you using FileVault, too?
I am indeed using FileVault.
I also am using FileVault.
Since the issue appears to be APFS + FileVault only, I wonder if a potential workaround might be:
1. Create a sparse bundle disk image with disk utility, and format it as HFS+
2. Move the VM files in this disk image and run from there
Assuming the issue is somewhere on Vmware’s side, this might trick it into working like it should.
I’ll try this when I get a chance - but if someone beats me to it, let us know if it helps!
UPDATE: tried this a bit after writing it; so far so good, 3GB Windows 10 VM resumes in about 30 seconds. However, it's too soon to celebrate, as using this same VM on APFS/FileVault had its resumes get (much) slower over the course of a few days, so I'll keep an eye on it for the next few days and let you know if it seems to be helping for real or not.
One more data point. Created a 200GB disk image, HFS+, not encrypted, residing on the same disk. Suspended my VM, copied the .vmwarevm bundle into the new volume, and started it from there. Resume was under 10 seconds.
So we have the very same host, the same VM, the same disk hardware.
I think we are quite sure now that it is related to APFS+FileVault.
At least we have some workaround now.
So far so good - I've resumed this VM 10's of times now over several days, and it's behaving as expected, with resuming being nice and fast.
I think we can cautiously declare victory - and hopefully this will provide enough clues that VMware can fix this!
I wonder if it's related to the new time machine snapshotting. It doesn't seem to work properly - wonder if it's backing up excluded files.