Mojave and External Monitors - Windows Server High CPU

[tommiy]

Every since upgrading to the last couple of Mojave supplemental updates my MAC has been suffering High CPU utilisation on the WindowServer process at the same times each day. Initially I thought this was the DisplayLink Drivers because when the event happens they jump to 150% CPU utilisation and WindowServer goes to 30-40%. Disconnecting the external monitors would rid the DisplayLink High CPU but the WindowServer would remain high. To clear the windowServer from running at 30-40% requires a dual reboot.

In case there was some thing incorrect I did a complete re-install of Mojave from Scratch....to further my testing.

Yesterday I ran without an external monitor and had no problems.

Today I uninstalled DisplayLink and ran with a Thunderbolt dock and 2 external monitors a old Apple 27 inch and a 4k LG.

Today, again at the same time the WindowServer process has shot up to 40% CPU utilisation. I'm at a lost as to how to identify the issue moving forward and am looking for help.

As I write this my WindowServer is now sitting at 52% with a static screen. Some where between 1520 and 15:42 today some thing happened that is obviously scheduled that upset the Mac. The system logs are unclear to myself exactly what that is that is causing the issue. Open to help please.


Sep 20 17:09:15 MAC iTunesCacheExtension[984]: objc[984]: Class ITNSImage is implemented in both /System/Library/Frameworks/iTunesLibrary.framework/Versions/A/iTunesLibrary (0x7fff95b942a0) and /Applications/iTunes.app/Contents/PlugIns/iTunesCacheExtension.appex/Contents/MacOS/iTunesCacheExtension (0x104dfd370). One of the two will be used. Which one is undefined.
Sep 20 17:10:02 MAC com.apple.xpc.launchd[1] (com.apple.WebKit.WebContent.96B532FB-FB2D-4D7D-BBAF-C286CE00FB56[988]): Service exited due to SIGKILL | sent by launchd[1]
Sep 20 17:10:42 MAC com.apple.xpc.launchd[1] (com.apple.WebKit.WebContent.99F33D5F-1DD2-4380-A996-2343192ABB72[989]): Service exited due to SIGKILL | sent by launchd[1]
Sep 20 17:10:45 MAC com.apple.xpc.launchd[1] (com.apple.WebKit.WebContent.47D026D8-E1C5-4755-8774-1EE3BBF36642[990]): Service exited due to SIGKILL | sent by launchd[1]
Sep 20 17:12:39 MAC com.apple.xpc.launchd[1] (com.apple.imfoundation.IMRemoteURLConnectionAgent): Unknown key for integer: _DirtyJetsamMemoryLimit
Sep 20 17:12:39 MAC com.apple.xpc.launchd[1] (com.apple.automator.xpc.runner.E6C13BA6-F33A-42BA-B3F6-F378F8AF4B49[991]): Service exited due to SIGKILL | sent by launchd[1]
Sep 20 17:13:11 MAC com.apple.xpc.launchd[1] (com.apple.imfoundation.IMRemoteURLConnectionAgent): Unknown key for integer: _DirtyJetsamMemoryLimit
Sep 20 17:13:11 MAC com.apple.xpc.launchd[1] (com.apple.automator.xpc.runner.2A13504E-7E00-43A4-8C34-9A2A5EC1D68E[995]): Service exited due to SIGKILL | sent by launchd[1]
Sep 20 17:13:14 MAC com.apple.xpc.launchd[1] (com.apple.imfoundation.IMRemoteURLConnectionAgent): Unknown key for integer: _DirtyJetsamMemoryLimit
Sep 20 17:13:14 MAC com.apple.xpc.launchd[1] (com.apple.automator.xpc.runner.78DFFC83-05F8-4AEB-A83C-3FC989895BDC[1000]): Service exited due to SIGKILL | sent by launchd[1]
Sep 20 17:15:35 MAC systemstats[54]: assertion failed: 18G95: systemstats + 673819 [D8A96B8F-0E65-3EA8-9EEE-982901402221]: 0x5
Sep 20 17:15:35 MAC syslogd[41]: ASL Sender Statistics
Sep 20 17:16:31 MAC com.apple.xpc.launchd[1] (com.apple.WebKit.Networking.6E385C76-0094-44C1-A5CD-488AA075EB54[1005]): Service exited with abnormal code: 1
Sep 20 17:23:06 MAC com.apple.xpc.launchd[1] (com.apple.WebKit.Networking.D878A7CC-19D9-4DB5-BD46-D717D9A0078F[1035]): Service exited with abnormal code: 1
Sep 20 17:25:35 MAC systemstats[54]: assertion failed: 18G95: systemstats + 673819 [D8A96B8F-0E65-3EA8-9EEE-982901402221]: 0x5
Sep 20 17:25:35 MAC syslogd[41]: ASL Sender Statistics
Sep 20 17:26:28 MAC com.apple.xpc.launchd[1] (com.apple.imfoundation.IMRemoteURLConnectionAgent): Unknown key for integer: _DirtyJetsamMemoryLimit
Sep 20 17:27:34 MAC com.apple.xpc.launchd[1] (com.apple.imfoundation.IMRemoteURLConnectionAgent): Unknown key for integer: _DirtyJetsamMemoryLimit
Sep 20 17:28:55 MAC com.apple.xpc.launchd[1] (com.apple.imfoundation.IMRemoteURLConnectionAgent): Unknown key for integer: _DirtyJetsamMemoryLimit
Sep 20 17:32:33 MAC com.apple.xpc.launchd[1] (com.apple.quicklook[1061]): Endpoint has been activated through legacy launch(3) APIs. Please switch to XPC or bootstrap_check_in(): com.apple.quicklook
Sep 20 17:32:38 MAC com.apple.xpc.launchd[1] (com.apple.imfoundation.IMRemoteURLConnectionAgent): Unknown key for integer: _DirtyJetsamMemoryLimit
Sep 20 17:32:42 MAC xpcproxy[1065]: libcoreservices: _dirhelper_userdir: 529: bootstrap_look_up returned (ipc/send) invalid destination port
Sep 20 17:32:42 MAC iTunesCacheExtension[1069]: objc[1069]: Class ITNSImage is implemented in both /System/Library/Frameworks/iTunesLibrary.framework/Versions/A/iTunesLibrary (0x7fff95b942a0) and /Applications/iTunes.app/Contents/PlugIns/iTunesCacheExtension.appex/Contents/MacOS/iTunesCacheExtension (0x108b9e370). One of the two will be used. Which one is undefined.
Sep 20 17:32:42 MAC com.apple.xpc.launchd[1] (com.apple.imfoundation.IMRemoteURLConnectionAgent): Unknown key for integer: _DirtyJetsamMemoryLimit
Sep 20 17:35:35 MAC systemstats[54]: assertion failed: 18G95: systemstats + 673819 [D8A96B8F-0E65-3EA8-9EEE-982901402221]: 0x5
Sep 20 17:35:35 MAC syslogd[41]: ASL Sender Statistics
Sep 20 17:45:45 MAC systemstats[54]: assertion failed: 18G95: systemstats + 673819 [D8A96B8F-0E65-3EA8-9EEE-982901402221]: 0x5
Sep 20 17:45:45 MAC syslogd[41]: ASL Sender Statistics
Sep 20 17:48:31 MAC fud[1082]: objc[1082]: Class MobileAssetUpdater is implemented in both /System/Library/PrivateFrameworks/MobileAccessoryUpdater.framework/Support/fud (0x10773c698) and /System/Library/AccessoryUpdaterBundles/ApplePowerAccessoryUpdater.bundle/Contents/MacOS/ApplePowerAccessoryUpdater (0x109af8a68). One of the two will be used. Which one is undefined.
Sep 20 17:48:33 MAC com.apple.xpc.launchd[1] (com.apple.quicklook[1084]): Endpoint has been activated through legacy launch(3) APIs. Please switch to XPC or bootstrap_check_in(): com.apple.quicklook
Sep 20 17:49:14 MAC diagnosticd[1091]: Posting empty stream filter
Sep 20 17:49:14 MAC diagnosticd[1091]: allowing Console (1090) access to stream due to admin status
Sep 20 17:49:14 MAC diagnosticd[1091]: send_stream_request_block_invoke: connecting to streaming server
Sep 20 17:49:14 MAC diagnosticd[1091]: Posting stream filter: "{
global = 30064771083;
}"

 

[tommiy]

My log above is also in the system report under Sync Services. I've no idea what that means but guessing the continued repeat of every thing being in 2 locations may not be good?

 

[tommiy]

wellp no ideas i trashed my MAC again. New install of Mojave and now wait till 17.30 to see if it goes silly with cpu utilisation on it. There is nothing else installed and i'd say by the difference in the log files the supplementary updates have potentially caused this issue.

 

[tommiy]

Sigh no one with any idea's is disappointing.

I'm at a loss to how to further fault find this. This is a now a cross post with Apple Forums but no response there either.

First, this issue only occurs if there is an external monitor plugged into my 2017 MAC at 1740. If there is no external monitor attached to the MAC there is no issue.

With an external monitor attached via USB-C/Thunderbolt 3 the WindowServer process goes High CPU at 1740 each and every day. It remains high CPU until exactly 30 minutes later at 1810. Rebooting between these times simply returns you to same position with the WindowServer Process running extremely high. Originally I thought this was display link drivers but I have proven that incorrect via uninstalling the driver and then only attaching a thunderbolt monitor to the MAC. Rules out Display Link.

I've completely reformatted the MAC, reinstalled Mojave 10.14.6 and reinstalled Office and a couple of other apps and this problem still happens.

I even have a video of this problem happening its that repetitive. Immediately prior to the CPU utilisation increase there is nothing visible in the activity monitor that would show why the increase is occurring. Similar, given I know exactly what time the issue is going to occur I know where in the logs to look and there appears to be nothing at all logged other than my bluetooth mouse.

I'm lost for idea's other than ensuring I disconnect the monitor prior to 1740!!! The only other option I can think of is returning to High Sierra to see if the problem remains. I don't see anyway of reverting back to other interim Mojave releases prior to 10.14.6.

Any suggestions ? I'd really like not to lose 30 minutes of my day while my MAC freaks out but I have run out of ideas.

Maybe High Sierra is the answer or setting a reminder everyday to unplug the monitors prior to 1740?

I'm really at a loss.

 

[Ritsuka]

I saw the same issue, WindowServer and others app consuming a bit of cpu only for 30 minutes after 18.00 when an external monitor is connected. But I can't say I found a solution. Weird issue.

 

[tommiy]

Thanks Ritsuka. Yes if you have Displaylink installed its more noticeable because Displaylink goes to some thing like %180 CPU utilisation while the WindowServer sits at 40%. In effect half my MAC being consumed doing nothing...

Looking through the log show --debug --start at 1739 I repeatedly see this being logged. The time span between the messages is probably the CPU increase I'm assuming. This logging of flipping the the display from 2880x1800 to 2880 x1800 continues until whatever is causing it finishes at 1810 each day.

I've no idea what is happening but the MAC gets so hot that its like a bar heater after that 30 minutes of running with Displaylink installed.

I'm not sure what kicks this off I can not see anything in the logs. The only thing occurring prior to this is advertisements of Bluetooth wireless.

So there is kernel logging about the Graphics Framebuffer that seems to range from 0.01 ms apart to 20-30ms apart. At 1810 this message goes away......


2019-09-24 17:39:26.011562+1000 0x2e8 Default 0x0 71 0 diskarbitrationd: [com.apple.DiskArbitration.diskarbitrationd:default] <private>

2019-09-24 17:39:26.038305+1000 0x71e Default 0x0 0 0 kernel: (AppleIntelKBLGraphicsFramebuffer) [IGFB][INFO ] [Flip_Transaction] FB0: Flipping display with wsaa 1 flipped from 2880 x 1800 -> 2880 x 1800 on 49000000 with 801 options at stride 11520

2019-09-24 17:39:26.038311+1000 0x71e Default 0x0 0 0 kernel: (AppleIntelKBLGraphicsFramebuffer) [IGFB][INFO ] [Flip_Transaction] Pipe Gamma Enable

2019-09-24 17:39:26.038312+1000 0x71e Default 0x0 0 0 kernel: (AppleIntelKBLGraphicsFramebuffer) [IGFB][INFO ] [Flip_Transaction] Disabling Render Compression

2019-09-24 17:39:26.084544+1000 0x71e Default 0x0 0 0 kernel: (AppleIntelKBLGraphicsFramebuffer) [IGFB][INFO ] [Flip_Transaction] FB0: Flipping display with wsaa 1 flipped from 2880 x 1800 -> 2880 x 1800 on 49000000 with 1 options at stride 11520

2019-09-24 17:39:26.118642+1000 0x71e Default 0x0 0 0 kernel: (AppleIntelKBLGraphicsFramebuffer) [IGFB][INFO ] [Flip_Transaction] FB0: Flipping display with wsaa 1 flipped from 2880 x 1800 -> 2880 x 1800 on 46600000 with 801 options at stride 11520

2019-09-24 17:39:26.118649+1000 0x71e Default 0x0 0 0 kernel: (AppleIntelKBLGraphicsFramebuffer) [IGFB][INFO ] [Flip_Transaction] Pipe Gamma Enable

2019-09-24 17:39:26.216588+1000 0x71e Default 0x0 0 0 kernel: (AppleIntelKBLGraphicsFramebuffer) [IGFB][INFO ] [Flip_Transaction] FB0: Flipping display with wsaa 1 flipped from 2880 x 1800 -> 2880 x 1800 on 46600000 with 1 options at stride 11520

2019-09-24 17:39:26.244896+1000 0x71e Default 0x0 0 0 kernel: (AppleIntelKBLGraphicsFramebuffer) [IGFB][INFO ] [Flip_Transaction] FB0: Flipping display with wsaa 1 flipped from 2880 x 1800 -> 2880 x 1800 on 49000000 with 801 options at stride 11520

2019-09-24 17:39:26.244900+1000 0x71e Default 0x0 0 0 kernel: (AppleIntelKBLGraphicsFramebuffer) [IGFB][INFO ] [Flip_Transaction] Pipe Gamma Enable

2019-09-24 17:39:26.351724+1000 0x71e Default 0x0 0 0 kernel: (AppleIntelKBLGraphicsFramebuffer) [IGFB][INFO ] [Flip_Transaction] FB0: Flipping display with wsaa 1 flipped from 2880 x 1800 -> 2880 x 1800 on 49000000 with 1 options at stride 11520

2019-09-24 17:39:26.378806+1000 0x71e Default 0x0 0 0 kernel: (AppleIntelKBLGraphicsFramebuffer) [IGFB][INFO ] [Flip_Transaction] FB0: Flipping display with wsaa 1 flipped from 2880 x 1800 -> 2880 x 1800 on 46600000 with 801 options at stride 11520

2019-09-24 17:39:26.378824+1000 0x71e Default 0x0 0

 

[tommiy]

well I contacted Apple regarding the kernel logging of constantly switching resolutions in10.14.6 18G95 and what causes that to occur. After a period of time it was suggested to disable night shift. Its only been a day but it is the first day where WindowServer has had run at high CPU. Appears that night shift is attempting to also control the external monitors and failing to do so for some reason. I have fingers crossed that this has fixed the immediate issue but will also be feeding back on my case so that Apple can fix Catalina (if it is fixed).

Noting that this appears to have originate from the first supplementary update to 10.14.6 for myself.

Hope this may assist some others.

 

[tommiy]

100% repeatable after sundown by turning nightshift on and off and WindowServer goes high CPU each time its enabled......confirmed with Apple. Now to wait for a Catalina fix.

 

[tommiy]

Thought I would try SU2 to see if it did anything to fix the breakage from the earlier SU. Nope. Still can cause WindowServer process to go high at the time of sunset and sunrise. Directly linked to enable/disable nightshift......have to keep waiting.

 

[baronmax]

Thanks so much for posting this @tommiy . Its May 2020 and I have exactly the same repeatable behaviour. Someone should be given the boot for such sloppy coding! The only thing to add is that if I turn on *manual nightshift* (until tomorrow) then windowserver behaves itself. I also have an HDMI connection to an external monitor (not displayport) with my 2015 Macbook pro. So glad I found this!

 

[matts1501]

Thank you @tommiy for the post and updates. This also just helped me with a 2018 Mini running 10.14.6. And like @baronmax experienced, windowserver seems to pipe down after manually toggling an already enabled nightshift from schedule.

 

[baronmax]

I don’t seem to have any problems any more. The big big difference with my illyama monitor (and maybe an unrelated problem, but maybe not) was to stop the auto switching on inputs on the monitor. After wake up from sleep this would cause my MacBook Pro 2015 no end of problems. Now after setting the monitor to just the hdmi input no problems... such an easy solution -grrr! Anyway hope this might help!

 

[tommiy]

Unfortunately while the high CPU problem caused by the kernel continously switching resolutions was solved in 10.15.1 it causes many other external monitor issues. All reported to Apple. I had one call with Apple engineers who were going to get back to me and then nothing for the next 8 months. I've simply not upgraded past 10.15.1 because of all the problems the subsequent updates have caused with external monitors in addition to my original issues. You can probably easily find those threads on apple communities they are the ones with over 85,000 view typically. I suppose I am saying if you can remain on Mojave. Do not try Catalina on an external drive as that upgrades your firmware which then causes problems back in Mojave. There is no way of returning to Mojave firmware. Its a one way street here and once you use the new firmware you seem to be broken. My current issues on 10.15.1 is that MACOS will get the wrong bit rate for the external monitor if i switch it. For HDMI it wants to go to HBR2 instead of HBR3 for some unknown reason. So i have to unplug the MAC and replug it in to get a full resolution external monitor. At least this is an easy work around. Those on later versions of Catalina have been left with no work around and purchasing new external hardware to try and get some thing that works. Moral to the story is turn off upgrades and do not trust Apple to deliver quality unfortunately.