secd chews up CPU cycles and Watch battery life goes in the terlet (2022) | Hacker News Reader

secd chews up CPU cycles and Watch battery life goes in the terlet (2022)

arstechnica.com

121 pointstosh31 comments

Random daemons running amok and using 100% CPU is a tale as old as Mac OS X. I've encountered them from time to time on every Mac I've had since I first installed 10.1 "Puma" on my lime green iMac DV in 2001. Sometimes you can just kill it and it'll stay dead, other times it'll respawn and continue its misbehavior. Sometimes a restart will permanently fix it, other times it'll happen occasionally until it just doesn't anymore. Often there are people asking about it on forums like this post, but rarely does anyone have any insight into what's going on.

There is a workaround described in the thread.

to every random daemon that macos 100% needs to run RIGHT NOW.

I don't get how the hell the company that is laser-focused on power efficiency - like to the point of having separate display scanout RAM just so they can ramp down system DRAM separately - also has a perpetual problem with out-of-control daemons eating the CPU. iOS and it's derivatives should have the same problems macOS has, but we can't see any of the performance data on an iPhone (unless you're jailbroken), so this probably just crops up as random complaints about hot iPhones with flat batteries.

This information is in a sysdiagnose fyi

Windows provides public debug symbols, so someone could potentially debug a process like this and get some idea of what it's doing based on the stacktrace, like Bruce Dawson is known to do [1]. Does OSX provide such debug symbols? A cursory internet search only finds information for developers compiling their own code for OSX, but maybe I'm searching for it wrong.

[1]: https://randomascii.wordpress.com/2017/07/09/24-core-cpu-and... for example.

going on 15 years of macOS/iOS dev:

I'm not sure if this is 100% equivalent, but, you can use Sample Process in Activity Monitor to get a stacktrace.

Generally, you expect Apple provides nothing, other than binary libraries you can link against and headers.

Of course, from there, there's decompilation etc., but I'd wager 1 in 1000 used that, even in the iOS early days where you had to work around the framework if you were serious about solving a problem.

Probably ~1 in 10,000 now Swift is predominant & harder to do runtime magic in, the platform API surfaces have stabilized relatively, and private API scanning is rigorous and serious.

On macOS we call it the "File a radar" tool.

Also known as “bug reporting event horizon”

Apples version of the Microsoft feedback hub, aka network attached /dev/null

fs_usage will often tell you what's going on just by looking at process and file names. sudo fs_usage -w

Nope! We debug like real men, with a decompiler and elbow grease.

Whoa, I just recently started having issues with my watch battery dying overnight when it used to last without any problems. Guessing it’s the same thing.

I have the exact same watch model as the poster and I’ve recently had it die within hours on random days when it usually lasts more than a day. Hopefully it’s the same thing and not some unrelated obscure issue…

I noticed this issue as well. secd consuming 20-60% of CPU all the time (how I noticed it in Activity Monitor) but even worse: writing GBs of data within seconds according to the "Disk" tab.

One theory is that it can be caused by device activation going wrong at some point. The solution from the forum using cloudkit CLI did not work for me.

Fix for me was deleting (scary) the directory that was named like a GUID and all the files within them which basically wipes your Keychain including local stuff like touch id config (and then fetches Keychain things again from iCloud but nothing which only existed locally like touch id). I don't know if there are other negative side-effects of removing the directory + contents though (!).

That said: I only noticed because of Activity Monitor. I still had about a day of battery life which speaks for the efficiency of Apple Silicon I guess.

That said 2: Now I notice "WindowServer" takes up about 15% CPU all the time. Is that usual behaviour? Or is it a side-effect of Activity Monitor updating its own window? Is there a way to get Activity Monitor like output on macOS via the CLI?

Somewhat ironically, Activity Monitor is not particularly efficient about drawing itself. You can use top to get similar information from the command line if you'd like.

Apple basically remote-bricked my 2yo Watch. Since December the battery drain is so bad that the Watch became useless. Lasts about 3 hours. There are several discussion threads [1,2] where many users have the same problem, even with the latest Ultra models.

Apple's leadership is so hypocritical it's disgusting. They make a cringe video [3] about environmental efforts while they waste so much energy and deteriorate so many batteries with their broken software.

[1] https://discussions.apple.com/thread/255347991 [2] https://discussions.apple.com/thread/255346006 [3] https://www.youtube.com/watch?v=QNv9PRDIhes

You can report a bug by typing applefeedback:// into Safari on the phone. (Protip.)

Or contact support or email the CEO, as you like.

1: it is a watch, not a phone.

2: they linked two issues, if Apple can't report that what makes you think your protip will help?

Apple's discussions website(what was linked to) is not how you get Apple's attention. Apple makes it pretty clear they don't do anything with the discussions website other than make it available for users to support each other.

Is it really “pretty clear” when people have been making the same mistake since the late 90’s/early 2000’s when the forum was launched?

do you want to fix the problem or just complain about it? mail the executives directly if you want the former. The latter is what the forum is for.

If you read the words Apple uses when they talk about it, yes. I can't help it that most people don't bother to read them. I'm sure Apple is fine with it though, as it lowers their support costs.

Even if you didn't bother to read the words, the fact that no-one from Apple ever says anything there should make it pretty clear if one ever thought about it.

> 1: it is a watch, not a phone.

And it's paired to a phone. I meant what I said.

> 2: they linked two issues, if Apple can't report that what makes you think your protip will help?

Man, nobody ever believes me on here when I say things I'm right about.

I used to submit a lot of radars (now Feedback). At some point there was very little effort from their side, no responses, annoying replies wanting to get everything served on a silver platter, just no appreciation of us devs. It’s a give and take. With such attitude I’m not willing to do free QA work.

As an example, I can write "notifications animation corner smoothing glitch" and honestly it's hard for me how one cannot know exactly what I mean. This bug has been introduced in iOS 16. Why would I waste my time making a screen recording etc. when something so obvious doesn't get picked up internally

> As an example, I can write "notifications animation corner smoothing glitch" and honestly it's hard for me how one cannot know exactly what I mean.

No idea what you're talking about, this isn't even a grammatically correct sentence. so I can't blame Apple for not reacting to such gibberish...

There are so many glitches with that screen I’m not sure which one you’re talking about.

I mean sometimes my keyboard doesn’t even show up on the login screen on iOS.

The Messages one where the share tray share to iMessage and then can’t send is infuriating.

And once in a blue moon desktop Messages glitches where typing on the keyboard is actually changing the text in the conversant’s most recent chat bubble. I have NEVER seen anything like it.

They are in denial that their OS frameworks are a disaster, and that they should have adopted a safer, managed more productive SDK and language for iOS UI apps.

Good thing I didn't tell you to do it then. I told the other guy to do it.

You should report a battery life issue, as that's much more severe than a UI glitch and usually easier to fix.

return the watch?

After 2 years? Not even European standard 2 year warranty applies anymore.

The physical watch is 2 years old, the software isn't.

I wonder if there's a legal argument that the warranty period resets after a software update.

Reply on news.ycombinator.com