Windows 10 VM pauses periodically, ignores mouse/keyboard

Dmitry@Parallels said:
Gentlemen,

My apologies for the delay in the investigation.
@PragmaticPraxis, according to the technical report your virtual machine is located on the external USB 2.0 disk (WD Passport). If yes, then we strongly recommend you move the VM to the internal Mac disk. USB 2.0 I/O bandwidth is not enough to handle VM disk operations properly.

@RodB, The Apple hypervisor was indeed added in Parallels Desktop version 12 as an optional (I would say even experimental) hypervisor driver. Apple hypervisor comes short of the following matters comparing to the Parallels hypervisor:
- Performance: slower on VM startup and shutdown
- Stability: may crash more frequently
- Functionality loss: no PMU, nested virtualization, thermal monitoring, energy profiling
Parallels Hypervisor is the best one.

Please try to reduce the number of CPUs to 2 and set 4096MB of vRAM to the virtual machine. Please let me know if this helps.
Click to expand...


Dmitry,
For what it's worth, I have been using external drives (USB2, USB3, FW800 & TB1 & TB2) for YEARS with my many test VMs, through many versions of Parallels and MacOS, and using both Windows 7 and Windows 10. Never had a SINGLE issue as a result. Until the latest build of Parallels. So, it certainly feels as if this advice is just more delay and obfuscation and general failure. Oh, and I have the same issues with my production VM that IS on the system drive, which is an SSD. So yeah, drive performance isn't the issue.
As for reducing the CPU count, what exactly did we pay for with the Pro version then, if it's so gimped it won't work unless we use it like the cheaper non pro version? I know, the best performing, most stable, most value adding Parallels is the one that isn't running, because I'm using VMWare now. Sorry to be so snarky, but jack shit for a meaningful response over multiple MONTHS is a little hard to swallow, for a "premium", "professional" product.
And yes, I still want my money back. Parallels Pro is a ripoff.
 
One other note -- may or may not be helpful. I originally reported that, before the pausing incidents began, I could always tell it was going to happen because I'd start experiencing a delay when clicking the mouse to move the text cursor -- specifically having to click twice before the text cursor moved to where the mouse cursor was placed. I've further noticed that this is true only after having entered text. That is, after having entered text, I experience behavior where I need to click once and then click again before the text cursor will move to where the mouse cursor is located. Once this has been done, I can click anywhere and the text cursor follows the mouse cursor until I again enter text. So, it is like it is "switching" modes or something -- allowing text entry or recognizing mouse cursor location. Kind of strange.
 
RodB said:
One other note -- may or may not be helpful. I originally reported that, before the pausing incidents began, I could always tell it was going to happen because I'd start experiencing a delay when clicking the mouse to move the text cursor -- specifically having to click twice before the text cursor moved to where the mouse cursor was placed. I've further noticed that this is true only after having entered text. That is, after having entered text, I experience behavior where I need to click once and then click again before the text cursor will move to where the mouse cursor is located. Once this has been done, I can click anywhere and the text cursor follows the mouse cursor until I again enter text. So, it is like it is "switching" modes or something -- allowing text entry or recognizing mouse cursor location. Kind of strange.
Click to expand...
Thank you for this detailed information, Rod. Would it be possible to arrange a remote session with one of our secondline support engineers? Is there a chance to catch this behavior during the remote?
 
RodB said:
One other note -- may or may not be helpful. I originally reported that, before the pausing incidents began, I could always tell it was going to happen because I'd start experiencing a delay when clicking the mouse to move the text cursor -- specifically having to click twice before the text cursor moved to where the mouse cursor was placed. I've further noticed that this is true only after having entered text. That is, after having entered text, I experience behavior where I need to click once and then click again before the text cursor will move to where the mouse cursor is located. Once this has been done, I can click anywhere and the text cursor follows the mouse cursor until I again enter text. So, it is like it is "switching" modes or something -- allowing text entry or recognizing mouse cursor location. Kind of strange.
Click to expand...
Thank you for this detailed information, Rod. Would it be possible to arrange a remote session with one of our secondline support engineer? Is there a chance to catch this behavior during the remote?
 
Dmitry@Parallels said:
Thank you for this detailed information, Rod. Would it be possible to arrange a remote session with one of our secondline support engineer? Is there a chance to catch this behavior during the remote?
Click to expand...
Yes, I think we could arrange a remote session. Here's some additional detail:
I found that it is not simply a double-click but a measurable delay between typing and when the mouse input is recognized. The delay is at least 1 second but less than 2 seconds but it appears to be the same amount every time. It does this regardless of the app, even in standard Notepad. If I wait two seconds (long enough for the delay to end) and then click the mouse, it immediately responds to the mouse input. If I begin clicking the mouse immediately after typing, I can click any number of times and there is no response until the delay ends.

I ran a test by opening Notepad in both my Win10 VM and my Win7 VM. In Win7, after typing a letter and then immediately clicking the mouse, the mouse input is immediately acted upon and the text cursor moves to the location of the mouse cursor. In Win10, after typing a letter, clicking that mouse as fast as I can does absolutely nothing until the delay has completed and then the text cursor moves to where the mouse cursor is located. Or, if I type a letter then simply wait two seconds, and then click the mouse, the text cursor immediately moves to where the mouse cursor is located.
 
Last edited:
Sorry for joining this discussion late, but I'm having the same issue. In my case mouse clicks stop working for 5 to 10 seconds although the cursor does move. Keyboard response is fine during the freeze. It's been happening ever since setting up Windows 10 probably 8 months ago or so at least 4 to 5 times per day. I never experienced this with Windows 7. I haven't seen any increase of CPU usage during these freezes, I'm not running the VM on an external drive and don't have the USB or bluetooth hardware enabled. I'm running only 2 cores, and my workload is pretty light. I'll monitor this thread in case I can add anything else, but it's good to know I'm not alone with this at least.
 
ThomasK7 said:
Sorry for joining this discussion late, but I'm having the same issue. In my case mouse clicks stop working for 5 to 10 seconds although the cursor does move. Keyboard response is fine during the freeze. It's been happening ever since setting up Windows 10 probably 8 months ago or so at least 4 to 5 times per day. I never experienced this with Windows 7. I haven't seen any increase of CPU usage during these freezes, I'm not running the VM on an external drive and don't have the USB or bluetooth hardware enabled. I'm running only 2 cores, and my workload is pretty light. I'll monitor this thread in case I can add anything else, but it's good to know I'm not alone with this at least.
Click to expand...
Hi @ThomasK7, please note the time when this behavior happens next time and send us a tech report mentioning the timings (Help > Send Technical Data...). Please post the ID here and I will take a look.
 
@Dmitry@Parallels: A couple months ago I sent some core dumps as requested. As I noted in the description, I wasn't sure if I was able to capture anything helpful due to how quickly the events occur and then cease. But I'd like to know if there was any follow-up on those as I've heard nothing back.
And, just to be clear, the issue continues to happen. It happened again just today.
 
@Dmitry@Parallels : Hello, anybody home? This continues to occur regularly. Did the core dumps provide any useful information? I was corresponding with an engineer on this but I do not remember his name. (Unfortunately, those emails have since been removed due to our corporate email policy expiration.)
 
ThomasK7 said:
Sorry for joining this discussion late, but I'm having the same issue. In my case mouse clicks stop working for 5 to 10 seconds although the cursor does move. Keyboard response is fine during the freeze. It's been happening ever since setting up Windows 10 probably 8 months ago or so at least 4 to 5 times per day. I never experienced this with Windows 7. I haven't seen any increase of CPU usage during these freezes, I'm not running the VM on an external drive and don't have the USB or bluetooth hardware enabled. I'm running only 2 cores, and my workload is pretty light. I'll monitor this thread in case I can add anything else, but it's good to know I'm not alone with this at least.
Click to expand...
I've got the same Problem and this problem exist a long time now, at least 2 or 3 Parallels Versions... I've tried everything like reducing CPU cores, set different boot flags, adjust mouse and keyboard settings etc. but the issue still remains. It's not as bad as when the problem emerged the first time, but it's still frustrating to move the mouse around until windows recatches it's position and I'm able to click something again. Meantime, Keyboard input always works, no issues there
 
@Dmitry@Parallels - Please respond and provide additional instructions as to information I can provide to help identify and resolve this defect.
@AndreasH10 - For me, it is good to know that it is not only me but I am sorry you are experiencing this; it is very frustrating -- and more frustrating that Parallels is apparently ignoring the issue (and us) now. I worked with engineering for a while -- sending them some core dumps and other info -- but since then I have heard absolutely nothing.
 
Interesting. I came to comment on the fact that I have not seen this issue since updating to 13.1.1, but I see no mention of it as one of the bugs fixed. And here I see others are still having the issue. Very frustrating that Parallels doesn't seem to be taking this seriously at all.
 
@PragmaticPraxis - Interesting. I am still running 12.2.1 but have heard nothing from Parallels about upgrading. When I check for upgrades, it says I have the latest. Yes, definitely very frustrating.
 
Hi,
Having the same problem here. At the beginning I thought it something chaotic but it keeps happening all the time and it is very disrupting and annoying. Any clue when it is gou=ing to be fixed (it somewhat puzzling that PD do does not gather automatically such misbehaviour and fix it on time)?

brand new MBP 15" i7 3.1Ghz 16GB running 10.13.2 Beta (17C60c)
 
Spoke too soon. The loss of input is back today. PD 13.1.1 on macOS 10.12.6 on mid 2014 MBP. Bit of a joke at this point, same problem across two versions and two years?
 
@Dmitry@Parallels - Just continuing to be a THORN IN PARALLELS SIDE. This continues to be an issue. It happens constantly, regularly, and I can now predict when it is going to happen as the system gets increasingly less responsive to changes between mouse clicks and text cursor placement until it finally does a pause, and then it starts all over again. Clearly, something is getting out of sync, filling up, something, that causes this to happen with Windows 10. I have received several updates but continue to be at version 1511.
 
@Dmitry@Parallels -- At least someone should reply back and provide an update, such as: "We continue to investigate this"; "We have no idea why this is happening and cannot duplicate it"; or "We are intentionally ignoring you, please go away".
 
@Dmitry@Parallels actually PMd me, claiming I was the only one having this problem, so the fact that it isn't fixed is because I haven't given them any actionable data. I pointed him at this very thread, and not a peep in a week. I don't think they are capable of fixing this.
 
@PragmaticPraxis - Wow. That does not say much about the consistency of support. I find that incredible given: 1) @Dmitry@Parallels and I interacted multiple times on this thread, and 2) as a result of that interaction someone called me from Parallels and I had several phone conversations with him in an effort to identify this issue further -- including my sending him/them multiple core-dumps.
 
RodB said:
@PragmaticPraxis - Wow. That does not say much about the consistency of support. I find that incredible given: 1) @Dmitry@Parallels and I interacted multiple times on this thread, and 2) as a result of that interaction someone called me from Parallels and I had several phone conversations with him in an effort to identify this issue further -- including my sending him/them multiple core-dumps.
Click to expand...

Directly from his PM
"Regarding the performance behavior on your Mac, we had never reviewed the tech data from your machine, the investigation just got frozen. I understand that it's not convenient at all to troubleshoot such behavior on your primary machine, but we can do nothing without analyzing tech details from your Mac. We don't have other users with this exact behavior, so we can't move further in investigation.
I have one though: is it theoretically possible that you move the virtual machine that experience performance issue from Passport to the Mac's internal HD and check it?
There is still a chance that all this mess happens due to the fact that your VMs are stored externally."

I've also told them a number of times I have exactly the same problem with VMs on the boot SSD, so the bullshit about the external drive is just wasted air. Kind of like their continuous suggestions to every new post suggesting following some procedure that we know know for going on a year doesn't work.
 
You must log in or register to reply here.
Back
Top