Chrome is Bad

Dec 12 2020
abruptly fib : Google Chrome installs an updater called Keystone on your calculator, which is bizarrely correlated to massive unexplained CPU use in WindowServer ( a system process ) [ 1 ], and made my whole computer slow even when Chrome wasn’t running. Deleting Chrome and Keystone made my calculator way, manner faster, all the time. Click hera for instructions. long history : I noticed my sword new 16 ” MacBook Pro started acting sluggishly doing even trivial things like scrolling. Activity Monitor showed *nothing* from Google using the CPU, but WindowServer was taking ~80 %, which is abnormally high ( it should use < 10 % normally ) . Doing all the normal things ( quitting apps, logging out other users, restarting, zapping PRAM/SMC, etc ) did nothing, then I remembered I had installed Chrome a while back to test a web site.

I deleted Chrome, and noticed Keystone while deleting some of Chrome ‘s other preferences and caches. I deleted everything from Google I could find, restarted the calculator, and it was like night-and-day. Everything was instantaneously and perceptibly faster, and WindowServer CPU was well under 10 % again .
then something else hit me, my family had been complaining about the sluggish operation of a 2015 iMac since much the day we bought it. I had tried everything I could think of – it had a Fusion drive and the symptoms were reproducible with a failing SSD – but drive diagnostics always turned up nothing. We even went deoxyadenosine monophosphate far as to completely wipe and set up the computer fresh multiple times .
then I remembered, installing Chrome was constantly one of the first things we did when we set up the calculator. I deleted Chrome, and all the files Keystone had littered on the computer, restarted, and it was indeed snappish it felt like a trade name new calculator. Yeah, I realize this sounds like a freakin ‘ infomercial, but it worked so well I spent a whole $ 5 on a sphere name and set up this web site even if it makes me sound like a rant en .

OK that’s weird, how do you delete Chrome and Keystone?

( Update 12/15/2020, the good people at Google on the Chromium team are taking this badly, if you ‘re technically disposed and can contribute samples please see this string – direct connection to the crbug 1158402. )

1. Go to your /Applications folder and drag Chrome to the Trash.
2. In the Finder click the Go menu (at the top of the screen), then click “Go to Folder...“.
3. Type in /Library and hit enter.
   • Check the following folders: LaunchAgents, LaunchDaemons, Application Support, Caches, Preferences.
   • Delete all the Google folders, and anything else that starts with com.google... and com.google.keystone...
4. Go to “Go to Folder...” again.
5. Type in ~/Library and hit enter. ( Note the “ ~ ” )
   • Check the following folders: LaunchAgents, Application Support, Caches, Preferences.
   • Delete all the Google folders, and anything else that starts with com.google... and com.google.keystone...
6. Empty the Trash, and restart your computer.

Now what browser should I use?

Safari is good and it ‘s already on your Mac. It ‘s fast and effective. If you need a Chromium-based browser, attempt Brave, Opera, or Vivaldi. ( Brave and Vivaldi both use an open source library called Sparkle for updates which makes an exit like this impossible. ) Firefox has pretty obtrusive pointer input reaction time which ( I, the author ) am pretty nitpicky about, but other than that it ‘s finely. ( Mozilla are a bunch together of short-sighted dopes for firing the Servo team. If the Servo team reorganize, I ‘d be inclined to recommend anything they make down the road ) .

What’s the deal with Keystone anyway?

Wired first reported on Keystone in 2009, when Google put it into Google Earth. It has a hanker history of crashing Macs by doing bizarre things that should n’t be necessary for auto-update software to function .
To all the effective people at Google who work on Chrome : something is going on between the code you ‘re writing and what is happening on people ‘s computers. I hope you can track it down and give us an honest postmortem.

update 12/15/2020 :
Collected anecdotes : hypertext transfer protocol : //twitter.com/lorenb/timelines/1338892756752732169

FAQ

Just restarting probably fixed it.

No, the write out persists across restarts, no indexing/applications running, etc, and is otherwise completely unexplained. Nothing obvious appears in Activity Monitor, and many people have meticulously tried to troubleshoot to no avail ( obviously including restarting ), and the only thing that immediately and conclusively addressed the consequence was removing Chrome & Keystone .

It’s the observer effect.

certain, Activity Monitor natural process itself uses CPU, then good looking for the issue might show slenderly elevated WindowServer activity. That is not what is happening here. The magnitude of the return is enormous, WindowServer CPU thrashing dwarf whatever Activity Monitor itself is doing ( no, using ‘top ‘ does n’t make an appreciable difference ). For many people actually measuring the remainder quantitatively is irrelevant ( though helpful ) since it ’ s so immediately obvious what the trouble was through vastly improved responsiveness, battery use, sports fan noise, etc .

This is a placebo!

The write out is n’t insidious, and the improvement is immediate, obvious, and testable. Please see above .

Hiding from Activity Monitor?! Inconceivable!

( The master translation of this page implied that keystone itself was hiding from Activity Monitor, my bad for the imprecise langauge. )
Correct, the keystone/updater process itself doesn ’ thymine hide itself from Activity Monitor, it briefly shows up and disappears on schedule. That is not the write out. It is causing something else on the system to consume massive CPU that leaves no indication that Chrome/Keystone are in fact the culprits .
certain, the spectrum of plausible theories here varies, anywhere from an feat trampolining through keystone, then code injecting WindowServer ( improbable, and insane if truthful … my gut says this issue is old, but still happens post-hardening ) to Google abusing system apis ( more likely ) to a bug in the OS that Google didn ’ thyroxine rigorously test against ( very possible ) .

I wouldn’t ascribe this to malice on Google’s part.

I agree, I personally believe the Chrome engineers and PMs who claim not to know about the consequence. That said, Google took on the duty of architecting an updater in a brittle and dangerous way when better / simple alternatives existed. good technology would make a problem like this structurally impossible.

If not malice, hubris, disparaging incredulity ( “ works on my car ” ), apathy, and hapless organizational prioritization would explain it .

You didn’t prove anything.

true. ( 12/18/2020 The original adaptation of this page was written in hurry but uncovered a substantial and systemic issue, that — had it affected person else — may have been written off wholly. ) This is going to be like proving whether smoking causes cancer. But whatever it is absolutely trash my computers ( along with many others ), and was fixed decisively after deleting Chrome & Keystone when nothing else worked .
On some level, the burden of proofread lies with Google to explain their architectual decisions since there is no technical reason for Chrome/keystone to run a haunting devil tied for people who do n’t use Chrome frequently, equitable to facilitate auto-updating ( which, for browsers, is a dependable theme ) when better solutions exist .