VMware Communities
cthree
Contributor
Contributor

Fusion hung on Leopard

My Fusion VM running CentOS5 on a Mac Pro running Leopard has hung.

Console reports:

27/10/07 1:55:38 PM vmware[1805] *** -[NSCFDictionary objectsForKeys:notFoundMarker:]: marker is nil - this will be an exception in applications linked on Leopard.

The vm is running CentOS5. I was compiling PHP 5.2.4. The VM is configured for 1 processor, 1GB RAM.

VMware Fusion is totally unresponsive and had to be killed. FYI.

Reply
0 Kudos
38 Replies
Danny_Joe
Contributor
Contributor

My new MacBook Pro (XP SP2 VM on Leopard with Fusion beta) was constantly hanging as well until I configured the Time Machine to exclude the VM folder.

Reply
0 Kudos
marjohan
Contributor
Contributor

I can confirm also, that filtering out the Virtual Machine -folder from Time Machine Backup plan/scan fixed the crashing, hanging and jamming problem of the Fusion in Leopard.

Reply
0 Kudos
cthree
Contributor
Contributor

I can confirm it does still crash for me even with TM off as I now have it set. I'm interested to see what hardware people are running it on. A colleague is using Fusion successfully with a MacBook Pro/Core2 Duo. However I'm using a MacPro:

2.66Ghz Mac Pro 2x2 Xeon w/4GB RAM

Are the any people here using Fusion + Leopard + Xeon Mac successfully? Any using Linux as the guest OS?

Reply
0 Kudos
Speedbird
Contributor
Contributor

Fusion was also becoming unresponsive forcing a "cold" boot in my Mac, all I did was exclude the folder containing my VMs (in my case are in "~/Documents/Virtual Machines"), after almost a whole day of crashing, read this thread's suggestion and applied the TM changes mentioned, my VMs have not crashed and are behaving normally again (Thank God!)

My relevant hardware specs: iMac C2D 2.1 (White iMac - Late 2006), 2Gb RAM

Software and VM OSes: Fusion Version 1.1rc1 (61385), Fedora Core 7, CentOS 4.5 and WinXP SP2

-- sb

Reply
0 Kudos
bflad
Enthusiast
Enthusiast

Are the any people here using Fusion + Leopard + Xeon Mac successfully? Any using Linux as the guest OS?

I am running this exact setup fine with multiple Linux VM's. (Ubuntu 7.04, Ubuntu 7.10, Fedora Core 7, etc.)... specifically I've compiled PHP just fine as well. I do notice that all of mine are running kernels 2.6.22+ for 64-bit.

~Brian

Reply
0 Kudos
mstocum
Contributor
Contributor

I have two Mac Pros here that both are choking with Leopard and VMWare 1.1rc1. Everything seems to work except for keyboard and mouse input. If I remote desktop in to the machine it functions perfectly. My MacBook Pro has no issues.

Reply
0 Kudos
admin
Immortal
Immortal

I have two Mac Pros here that both are choking with Leopard and VMWare 1.1rc1. Everything seems to work except for keyboard and mouse input.

Output of `ioreg -l -w 0 | grep SecureInput` ?

Reply
0 Kudos
ssibert
Contributor
Contributor

I have a dual quad-core and I'm running Fedora7 x86_64 1cpu 384MB in a VM and XPSP2 2cpu 448MB in another VM. They were doing fine (updated vmtools in each VM) and then they hung. Spinning beach ball. I was doing lots of activity ("yum update" in the Fedora7 VM) and I think Time Machine was backing up. BUT I have always excluded the Fusion folder (where the VM's reside) from Time Machine. I have not excluded the Fusion application.

I noticed something at the end of my fusion log file. I'm not sure when the beach ball hang started but it could've been around the same time these messages were logged (all times are EDT):

Nov 01 10:16:36.895: vmfusion| VMHS: Connection to VM broken: cfg: /Volumes/HD-XP/Fusion/Fedora 7.vmwarevm/Fedora 7.vmx; error: Pipe connection has been broken; state: 3

Nov 01 10:16:36.895: vmfusion| Unmounting the vm: /Volumes/HD-XP/Fusion/Fedora 7.vmwarevm/Fedora 7.vmx

Nov 01 10:16:36.896: vmfusion| Unmounting VM complete: /Volumes/HD-XP/Fusion/Fedora 7.vmwarevm/Fedora 7.vmx

Nov 01 10:16:36.907: vmfusion| Mount state values have changed: /Volumes/HD-XP/Fusion/Fedora 7.vmwarevm/Fedora 7.vmx

Nov 01 10:16:36.908: vmfusion| Reloading config state: /Volumes/HD-XP/Fusion/Fedora 7.vmwarevm/Fedora 7.vmx

My Time Machine Prefs page shows the latest backup at 10:36am; I just installed Leopard yesterday afternoon so I don't know if that is the time TM started or ended its backup operation. The next backup is showing to happen at 11:31am.

Reply
0 Kudos
gazhay
Contributor
Contributor

Mine doesn't have any output.

VMFusion stalls with spinning ball and time machine enabled

Reply
0 Kudos
admin
Immortal
Immortal

Time Machine does not play well with a bunch of things, including virtual machines. You should exclude your virtual machine directory from being backed up.

Reply
0 Kudos
cshware
Contributor
Contributor

I have a Mac Pro, 2.66GHz model, running Leopard and VMware Fusion 1.1RC.

I had two VMs running: FreeBSD 6.x and FreeBSD 7.0beta.

I was logged into both machines with ssh, and the 7.x system stopped responding. The VMware GUI hung like other described here.

I could still use my ssh logins in the FreeBSD 6.x system.

I shut FreeBSD 6.x down but 7.x died with the ssh logins dying, and the 7.x VMware window going blank.

I eventually gave up waiting and force killed vmware, but it took several tries. I had to manually kill off the natd processes and other helpers.

Now, I have two processes stuck in state E (trying to exit) named vmware and vmware-vmx. I suspect the latter is the vmx running FreeBSD 7.x.

I run Time Machine, but the first thing I did when installing Time Machine was disable it for anything like virtualization, Aperature, and others where people report having problems. It's not even that good a system anyway.

I don't think Time Machine is the overall problem anyway.

If I continue to have issues after I reboot I'll try to report back here.

Reply
0 Kudos
bapper
Contributor
Contributor

cthree, did you ever get things working? I'm seeing exactly the same thing you are with various linux flavors running in the VM, time machine off under Fusion v1.1. I spent several days trying to get SOMETHING working to no avail. I may have to go back to Tiger if there isn't a resolution to this one already. I'm on a Macbook Pro Core 2, I seem to see the issue when doing disk intensive things like installing a new OS on a VM or compiling large apps. Of course, it has completely corrupted a few of my VMs at this point and I can't install new ones because of this issue.

Reply
0 Kudos
cthree
Contributor
Contributor

I never got it working any better than I described. I gave up having no more time to burn. Fusion under Leopard running Linux as a Guest OS never proved stable enough to be useful for me. I'm still waiting on some sort of solution and maybe I'll try again someday but deadlines and production schedules preclude goofing around with it any more right now. Maybe when 1.2 comes out I'll give it another look.

I had hoped to implement a VMWare virtualized environment across my entire server farm, running Fusion locally to allow me to test changes to my application server software/configuration before distributing a single, uniform VM image to my server farm (like installing and testing new server software and updates), much in the same way I test my application code before checking it in and deploying it to live servers. As it stands such and arrangement is too involved and complicated to be successful. IF I can get an image to run on my workstation such that I can make changes to it and test it then it might make sense to look at running VMware on my servers and deploying a tested image to them. We aren't there yet.

It would also be a huge benefit if I could manage that server farm with a Mac OS X native client application rather than having to run the Windows client inside yet another VM. That's not really the sticking point though.

Reply
0 Kudos
bapper
Contributor
Contributor

Guess replying via email didn't work. Smiley Sad

On a hunch I downgraded Fusion to 1.0 and have had no problems yet today. This is definitely an issue with 1.1 and not Leopard (or perhaps 1.1 in combination with Leopard). Hopefully this gets resolved with the next version. I'd be willing to try pre-release or debug code to try to resolve this if vmware does anything like that. I didn't see any way to really get feedback on the support page without paying money for support tickets, which I'm not willing to do on something that's obviously a regression.

Reply
0 Kudos
mcdavis
Contributor
Contributor

I started a thread about the colorwheel of death. Had those problems and...in my case...they were linked to an incompatibility with Filevault. At least that is the current theory.

DONT USE FILEVAULT on LEOPARD WITH FUSION running VISTA. It could be more generic: like don't ever run filevault...add to the list...with timemachine.

Reply
0 Kudos
bapper
Contributor
Contributor

DONT USE FILEVAULT on LEOPARD WITH FUSION running VISTA. It could be more generic: like don't ever run filevault...add to the list...with timemachine.

I think it's a far more generic problem, I don't own a copy of windows and use Fusion for linux VMs. I really don't have anything running other than the VM and a couple of terminals and the issue is ever present. It doesn't appear to be as bad with 1.0 but I still hit it while doing a yum update of a fresh F8 install. I don't run windows and run text console VMs (no graphics). I seem to see it with heavy disk/cpu activity (and possibly heavy network activity) in the VM and nothing else really going on at the time. Which is a bummer because I use this to do linux development work on my Mac. I'm hoping to at least get some sort of log output one of these times. I guess it is time to stress test the latest Parallels to see if it has the same problem.

Reply
0 Kudos
cthree
Contributor
Contributor

It's a real issue and it's not TimeMachine or FileVault or XXX related but rather something more fundamental. I think we can all agree that it is disk related and that it occurs wheren there is a significant amount of disk I/O but beyond that is for VMW to figure out.

FWIW I had the same problem on 1.0 which is why I switched to 1.1rc in the first place.

Reply
0 Kudos
shakabuku
Contributor
Contributor

I, too, am having the same problem. I'm using 10.5.1 on a second gen Macbook Pro w/3GB RAM. Filevault's not being used and I turned off Time Machine. I've been trying to install Fedora Core 6 for some upgrade run-throughs, but every time, Fusion (1.1, btw) beach-balls. The only way I can kill Fusion after that is by holding down the power button and forcing a reboot. All known clean shutdown methods hang (afaict).

Hopefully, this can get fixed up really soon.

Reply
0 Kudos
jared_oberhaus
Contributor
Contributor

The problems discussed in this thread are likely the same as that discussed in these threads:

Fusion 1.1RC1 "beachballs" under heavy VM load (after switch to leopard)

http://communities.vmware.com/message/820392

Hangs & Disk Corruption using Ubuntu 7.10 w/Fusion 1.1 + 10.5.1

http://communities.vmware.com/thread/115466

Reply
0 Kudos