launchd has written 40GB in 7hrs?

[thebart]

I was looking at system monitor, b/c well it's what I do with my new toy, and notice I have 45GB written to disk. I looked in system info and found it's only been 7 hours since last boot. That's quite a bit of data written in 7hrs. Turns out 40GB of that is by launchd. That's the launch daemon, the thing that launches things? What could it possibly be writing that much? Yikes

Mini M1 16/512 Ventura 13.2.1

 

[Fishrrman]

"What could it possibly be writing that much?"

Virtual memory disk swapping.
Apple changed how RAM and VM is handled some years back (around the introduction of "Mavericks").
Then, they made it much more "disk intense" with the introduction of the m-series CPU and "unified memory".

This is why there's a long-running thread in the Apple Silicon sub forum regarding extremely high rates of disk writes with the m-series CPUs.

There's a way to fix this, I've found.
I've posted about it numerous times, but very VERY few seem willing to even try it.

The solution is:
TURN OFF VM disk swapping, using the terminal.

Of course, you have to have enough RAM to do this.
8gb may not be "enough" -- but 16gb IS enough.
And you have to be willing to monitor application usage, closing un-needed apps.

Now, I have no VM disk swapping, because my Mac runs only in "real" memory.
No "virtual" memory... AT ALL...!

 

[thebart]

I normally have zero swap. The only thing I recall doing yesterday that might've caused swap was running 3dmark wild life. Could that be it?

Speaking of which: I only get a score of ~5000 running 3dmark wild life extreme on plain M1. iPad pro is 9000+. Is that right? That seems low.

 

[bogdanw]

It’s a good point about virtual memory, but on MBA M1 with only 8GB RAM I see
sysctl vm.swapusage
vm.swapusage: total = 0.00M used = 0.00M free = 0.00M (encrypted)
with Photoshop, iMovie, LibreOffice, Safari and Chrome opened.

 

[thebart]

It’s a good point about virtual memory, but on MBA M1 with only 8GB RAM I see
sysctl vm.swapusage
vm.swapusage: total = 0.00M used = 0.00M free = 0.00M (encrypted)
with Photoshop, iMovie, LibreOffice, Safari and Chrome opened.

I just checked and mine says 0.0 across the board. But what is the timeframe on that? Right now or since boot or something else.

 

[Brian33]

Turns out 40GB of that is by launchd

Virtual memory disk swapping

I seriously doubt that the launchd process is responsible for virtual memory swapping.

 

[bogdanw]

I just checked and mine says 0.0 across the board. But what is the timeframe on that? Right now or since boot or something else.

I presume it's right now. There is another command for average history, but I have no idea what the returned values mean. From the manual
"Information about the load average history may be obtained with:
sysctl vm.loadavg"

I seriously doubt that the launchd process is responsible for virtual memory swapping.

launchd is responsible for everything
" launchd – System wide and per-user daemon/agent manager
launchd manages processes, both for the system as a whole and for individual users."

 

[thebart]

Oh, one thing I just thought of. I had a kernel panic yesterday. Mac spontaneously rebooted then on startup asked if I wanted to send a report. First time that has happened, but I've only had it for 3 weeks. I didn't think to sve the panic screen.

 

[bogdanw]

Save the report if it happens again, the text of it.

Apple: "If your Mac restarts and a message appears" https://support.apple.com/guide/mac-help/if-your-mac-restarts-and-a-message-appears-mchlp2890/mac

 

[thebart]

I presume it's right now. There is another command for average history, but I have no idea what the returned values mean. From the manual
"Information about the load average history may be obtained with:
sysctl vm.loadavg"

Any idea how to interpret this?

sysctl vm.loadavg
vm.loadavg: { 1.34 1.35 1.34 }

 

[Brian33]

Any idea how to interpret this?

sysctl vm.loadavg
vm.loadavg: { 1.34 1.35 1.34 }

From 'man top': "LoadAvg Load average over 1, 5, and 15 minutes. The load average is the average number of jobs in the run queue."

Try the 'top' command -- see if you get the same numbers. I think you should.