ssd swap - high usage of Terabytes Written

[Mike Boreham]

turning off power nap and disabling swap when the computer sleeps is good, like the above poster mentioned.

sudo pmset -a powernap 0

Googling "Powernap and M1 Macs" suggests M1 Macs don't have powernap and it is not mentioned in the Sys Prefs anywhere.

Perhaps it would be more accurate to say there are no Powernap options in the user interface on M1 Macs because it is all built in on M1 Macs? Why would Apple make this change? What consequences of turning off, with pmset?

 

[Thistle41]

As the others above have mentioned, check your pmset settings. I found mine got reset when I updated to 11.4 due to the hibernation mode pmset setting being added. I have powernap on 0 like the others here suggest also.

My pmset settings are as follows:

System-wide power settings:


Currently in use:
standby 0
Sleep On Power Button 1
SleepServices 0
hibernatefile /var/vm/sleepimage
powernap 0
networkoversleep 0
disksleep 10
sleep 1
hibernatemode 0
ttyskeepawake 0
displaysleep 10
tcpkeepalive 1
womp 0

Thanks to all for the tips. I tried just the 'powernap = 0' last night and get improved results of around 1/2 what I had before:

9 June 22:14	10/06/2021 08:19	TBW
10/6/21 22:00	11/6/21 08:44
54,824,414	56,463,465	0.839194112

I will set the remainder as you describe @TheSynchronizer and log the results again.

 

[VitoBotta]

I haven't written in this thread in a while. I'm on Monterey now and the problem seems to be gone completely, I hope.
For reference, I'm at 268 terabytes written in 6 months, and the writes were adding up to hundreds of GB or even terabytes per day. I updated to Monterey one and half day ago and just 22 GB have been written total so far!

I am really hoping this is fixed for good now because my case was one of the bad ones.

 

[k-hawinkler]

I haven't written in this thread in a while. I'm on Monterey now and the problem seems to be gone completely, I hope.
For reference, I'm at 268 terabytes written in 6 months, and the writes were adding up to hundreds of GB or even terabytes per day. I updated to Monterey one and half day ago and just 22 GB have been written total so far!

I am really hoping this is fixed for good now because my case was one of the bad ones.

Many thanks. How stable is Monterey?
My M1 Mac mini on BS 11.4 occasionally still crashes during a YouTube videos.

 

[jdb8167]

I haven't written in this thread in a while. I'm on Monterey now and the problem seems to be gone completely, I hope.
For reference, I'm at 268 terabytes written in 6 months, and the writes were adding up to hundreds of GB or even terabytes per day. I updated to Monterey one and half day ago and just 22 GB have been written total so far!

I am really hoping this is fixed for good now because my case was one of the bad ones.

Did you see any drastic change in percentage used?

I’m trying to verify that Apple didn’t make any changes in reporting of SMART data as reported by a dubious article.

 

[VitoBotta]

Many thanks. How stable is Monterey?
My M1 Mac mini on BS 11.4 occasionally still crashes during a YouTube videos.

It looks pretty stable. The only problem I had was that when waking the mini from sleep it panicked and the magic mouse/keyboard got disconnected, so for now I am just keeping the computer always awake and just turning the display off. Other than that, rock solid so far.

Did you see any drastic change in percentage used?

I’m trying to verify that Apple didn’t make any changes in reporting of SMART data as reported by a dubious article.

Not at all. I don't believe that it was just a reporting issue.

 

[jdb8167]

Not at all. I don't believe that it was just a reporting issue.

Thanks. Can you tell us your percentage used and your drive size?

 

[VitoBotta]

Thanks. Can you tell us your percentage used and your drive size?

18%, 256GB

 

[jdb8167]

18%, 256GB

👍

 

[Mike Boreham]

18%, 256GB

That's another very useful one for the list of Mac Analytics Data reported results in this thread:

256GB Macs
Souko 1% 15.3TBW
Leons 1% 14.9TBW
The Synchroniser 2% 25.2TBW
VitoBotta 18% 268TBW

512GB Macs
Mike Boreham 0% 21.7TBW
Formalhaut 1% 25.9TBW

1TB Macs
jdb8167 0% 10TBW
Dieselm 6% 150TBW

Looks a reasonably consistent picture emerging. Factors like write amplification mention by @jdb8167 affect.

 

[Thistle41]

That's another very useful one for the list of Mac Analytics Data reported results in this thread:

256GB Macs
Souko 1% 15.3TBW
Leons 1% 14.9TBW
The Synchroniser 2% 25.2TBW
VitoBotta 18% 268TBW

512GB Macs
Mike Boreham 0% 21.7TBW
Formalhaut 1% 25.9TBW

1TB Macs
jdb8167 0% 10TBW
Dieselm 6% 150TBW

Looks a reasonably consistent picture emerging. Factors like write amplification mention by @jdb8167 affect.

Just to add mine to the above as a 8/256 MBA:
1% 29.7TBW.

Not sure why others with less TBW value have flipped over the 2% boundary?

An update on the sleep writing front. I tried setting those flags to '0' in pmset but some will not change. I trialed simply switching off over night and the cold start is fairly quick but you still end up with a bundle of SSD writes - 0.406652416 TBW in my case so a small real from the 'pmset -a powernap 0' which gave me 0.839194112.

I will look at installing Monterey beta 2 if it looks stable enough.

 

[Mike Boreham]

Just to add mine to the above as a 8/256 MBA:
1% 29.7TBW.

Not sure why others with less TBW value have flipped over the 2% boundary?

Thanks for the info. Yes yours does look a bit anomalous, but I don't think this is an exact science, with other factors affecting linearity.

Updated list:

256GB Macs
Souko 1% 15.3TBW
Leons 1% 14.9TBW
Thistle41 1% 29.7TBW
The Synchroniser 2% 25.2TBW
VitoBotta 18% 268TBW

512GB Macs
Mike Boreham 0% 21.7TBW
Formalhaut 1% 25.9TBW

1TB Macs
jdb8167 0% 10TBW
Dieselm 6% 150TBW

 

[Mike Boreham]

I haven't written in this thread in a while. I'm on Monterey now and the problem seems to be gone completely, I hope.
For reference, I'm at 268 terabytes written in 6 months, and the writes were adding up to hundreds of GB or even terabytes per day. I updated to Monterey one and half day ago and just 22 GB have been written total so far!

I am really hoping this is fixed for good now because my case was one of the bad ones.

@VitoBotta Thanks. Do you have any idea why your usage is so high? It seems an order or magnitude higher than most other people, although I know there are some others like you.

Such a difference ought to have an identifiable and unusual difference in usage.

Very good to hear that it has stopped for you in Monterey, presumably without any change in usage.

My own logging shows quite a daily variation without obvious reasons. I calculate GB/h and see variation between 1GB/h and 16GB/h, without an obvious change at 11.4.

 

[VitoBotta]

@VitoBotta Thanks. Do you have any idea why your usage is so high? It seems an order or magnitude higher than most other people, although I know there are some others like you.

Such a difference ought to have an identifiable and unusual difference in usage.

Very good to hear that it has stopped for you in Monterey, presumably without any change in usage.

My own logging shows quite a daily variation without obvious reasons. I calculate GB/h and see variation between 1GB/h and 16GB/h, without an obvious change at 11.4.

Unfortunately I haven't been able to pin point some specific app or apps that can cause the issue, and eventually I gave up after trying many things. For me the important thing is that now it's fixed apparently. I have rebooted 2 days ago and kernel_task has written only 128GB since then, which is awesome.

 

[Mike Boreham]

Unfortunately I haven't been able to pin point some specific app or apps that can cause the issue, and eventually I gave up after trying many things. For me the important thing is that now it's fixed apparently. I have rebooted 2 days ago and kernel_task has written only 128GB since then, which is awesome.

Thanks, but puzzling!

 

[Thistle41]

Well, I'm not sure what did this but since I left it in 'sleep' and went out at 12:00 today and looking at it now 20:33 I have used only 0.000246784TBW (previously I was getting hundreds of gb) which is pretty small. I think turning off the pmset sleep flag is probably the clue. So thanks for the reminder as I was considering the Beta version that has issues of its own.

 

[k-hawinkler]

What are the issues of the Beta version? TIA.

 

[Thistle41]

What are the issues of the Beta version? TIA.

Over on another thread there is a discussion about the beta build and ppl have reports of various things not working. Additionally it is a very large image to download (debug code I guess) so I will wait for now.

On the ssd topic, just checking again today there’s a sudden increase in writing for some unknown reason. Going to see what happens overnight.

Update edit:
This is the overnight so called 'sleep' useage.

58,821,043
59,635,123	0.41680896

21:47...7:47 = 10 hrs it wrote 416.8gb. What the hell was it doing?

 

[jdb8167]

Thanks, but puzzling!

There was a tweet by a very smart kernel hacker named Hector Martin @marcan42 (one of the people porting Linux to the M1) who said it was a kernel bug. Obviously it isn't something that affects everyone which probably makes it harder to track down. He was going to diff the kernel code once Apple updates its open source repository. We'll probably find out then.


Edit: Tweets

https://twitter.com/i/web/status/1396374313591140357

https://twitter.com/i/web/status/1399283505016754181

 

[Mike Boreham]

Someone else has pointed out that 16.1 TB (Terabytes) is close to 14.7 TiB (Tebibytes) so maybe DriveDx should be reporting 14.7 TiB.

TB to TiB conversion

Todays update to DriveDx corrects this.....without admitting that they have been incorrectly reporting TiB but labelling it TB.

 

[eldho]

I might be missing something here but is there a way to work out GB written from the statistics here? There is a number given for the data units written, but what does this mean?

 

[Ningj]

759068 * 512 / 1000 / 1000 / 1000 = 0.389TB (388.643GB)

 

[Maximara]

The SSD wear issue that has been on website for a long was brought up in icavedave's CANCELLED! The APPLE Products that never were! at the 17:48 mark. His position is effectively that, outside of very weird special cases, this is a tempest in a teapot. Also even though he doesn't read it "Crosschecking with Activity Monitor has indicated these reading by smartmontools seem to be correct." is in the post.

 

[kubilaydem]

Hey!

How about to use safari now? Is it ok now? Or should i continue to use firefox / brave.

 

[k-hawinkler]

I never had a problem with Safari, so have been using it.