WindowServer Craziness - what's the FIX, not the WORKAROUND?

I've already read several threads about this issue, some going back over a year; which is extremely disappointing, since what is such a serious bug, once it hits you, appears to have not been fixed in all that time.

Same issue here as the others report: starting Parallels Desktop 3 causes the windowserver process to peg out one of my CPU cores (Mac Pro Quadcore), and the entire system becomes essentially frozen. The only way to graciously get out of this problem in any reasonable amount of time is to SSH into the Mac Pro and use the "killall Parallels" command.

What's especially frustrating to me is that this just happened, out of the blue, for no apparent reason. Never an issue with Parallels starting on this machine in the past. In fact, I just switched back to Parallels for most of my usage a couple of months ago because I preferred it to VMWare.

Then, yesterday, a reboot required and all of a sudden Parallels won't start without killing the machine. I didn't upgrade any programs at all, no configuration changes, nothing but reboot the machine.

I tried reinstalling Parallels, but to no avail. I did find that I can get Parallels started by turning off my ARD (Apple Remote Desktop) and VNC applications, but this is completely unacceptable. People who run VNC or ARD services do it because they need remote access to the machines. What if I need to start a Parallels session from offsite for some reason?

So, while disabling VNC or any of the several other suggestions that get some people working might be a fine TEMPORARY workaround, I can't believer that for a year or more it's been the only solution offered. What's the FIX? When's it coming? Are you guys even working on it, or is this going to continue to be unaddressed? Any idea why an otherwise working machine would all of a sudden start experiencing this bug?

 

We are aware about this problem, and working to resolve this as soon as possible,
Currently in Parallels Server there are no problems, and in beta Parallels Desktop version 4
but PD4 and PS engine is different, so we are looking to find solution for PD 3

 

Thanks for the reply.

My biggest concern is how can this just happen out of the blue with no obvious changes to the system. Will my other Macs start displaying this problem for no apparent reason at all?

It's very strange to just boot up, after no system changes, and have a program that's been working for months (or longer), suddenly kill the entire system just by starting it up.

I'm the type who normally purchases upgrades and even goes beta once their stable enough for general production work, but I hate to feel like I have to use a beta of a new version of a product to get stability that I thought the original product had...

 

I understand situation, and believe me we are trying our best to fix all old problems, and introduce new product with best quality

 

Has there been any progress in addressing this? It went away for some reason, and is back again for me. it is mind-boggling that there is no solution for this after all this time. I can go months without an issue, and then all of a sudden I can only start Parallels if I kill ARD first. Still unacceptable.

Edit: I wonder if using SPACES feature might impact that. I have had Parallels set to auto-start into a different Spaces desktop. I've noticed an odd bug all along with Parallels and Spaces that when you try to select one of multiple machines on startup of Parallels that you can't click the select button, and trying to bounces you back and forth to Space 1 and the Parallels Space. The only way to select the machine to get it started is to hit the space bar...makes me think that with Window Server UI being the process that Parallels runs up, that Parallels is killing the machine by trying to bounce back and forth between Spaces on some sub-level while starting up. Possibility?

Anyways, I will try letting Parallels start up in the primary Space for a while and see if this helps the problem.

 

Nope. Spaces didn't stop this from happening. Whatever Space Parallels is set to start in, ARD running kills it. i don't know how I got it working before after my original post, but I'd say the recent update to ARD by Apple this week certainly seems a bit too coincidental to the problem resurfacing.

 

Do you have lines like "IPI stat: rescheduled X CPUs out of Y" in your system.log? If so, this a known problem and it will be fixed in the next update?