Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
ssd swap - high usage of Terabytes Written
- Thread starter Forti
- WikiPost WikiPost
- Start date
- Sort by reaction score
- Status
For me the biggest culprit is launchd - if Activity Monitor is to be believed, then so far in roughly 12 hours today, launchd has written 112GB, so that's about 9GB/hour. the kernel is a distant second place with 17GB (1.5GB/hour)
This is on a Mini M1 with 16GB memory and memory pressure always green. Main apps both Intel, both Adobe, Lightroom Classic and Photoshop.
This is on a Mini M1 with 16GB memory and memory pressure always green. Main apps both Intel, both Adobe, Lightroom Classic and Photoshop.
Hi guys, too many pages to read so apologies in advance...
I have had a mini for 3 months now and this is the situation:
Since the last reboot a few days ago 7.37TB have written to disk. In particular, kernel_task has written 5.17TB and the Nextcloud client 1.38TB. I will check with the Nextcloud communities what the client is doing on the main drive because its data folder is on an external drive. Unbelievable.
What can explain kernel_task's behavior? Again sorry if I haven't read the whole thread.... has anyone found any culprit?
I have had a mini for 3 months now and this is the situation:
Code:
=== START OF SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED
SMART/Health Information (NVMe Log 0x02)
Critical Warning: 0x00
Temperature: 37 Celsius
Available Spare: 100%
Available Spare Threshold: 99%
Percentage Used: 10%
Data Units Read: 300,657,713 [153 TB]
Data Units Written: 289,063,316 [148 TB]
Host Read Commands: 1,285,073,683
Host Write Commands: 787,048,167
Controller Busy Time: 0
Power Cycles: 71
Power On Hours: 693
Unsafe Shutdowns: 7
Media and Data Integrity Errors: 0
Error Information Log Entries: 0Since the last reboot a few days ago 7.37TB have written to disk. In particular, kernel_task has written 5.17TB and the Nextcloud client 1.38TB. I will check with the Nextcloud communities what the client is doing on the main drive because its data folder is on an external drive. Unbelievable.
What can explain kernel_task's behavior? Again sorry if I haven't read the whole thread.... has anyone found any culprit?
Kernel Task does a lot of things so its hard to say what's causing it. Is the client running through Rosetta?
Today I turned off spotlight and repeated basically the same workflow as Tuesday.
1.7 TB written in 6.5 hours, none of it my data. This is pretty much the same as yesterday so disabling spotlight makes no difference for me (well, apart from taking away a useful aspect of the operating system).
Tomorrow I'll try narrowing the apps I use even further to see its any one in particular that's triggering the really excessive usage.
1.7 TB written in 6.5 hours, none of it my data. This is pretty much the same as yesterday so disabling spotlight makes no difference for me (well, apart from taking away a useful aspect of the operating system).
Tomorrow I'll try narrowing the apps I use even further to see its any one in particular that's triggering the really excessive usage.
Then please leave forum and stop buying Apple. If Apple's hardware really "sucks" then don't buy it. More people have success than failure. I and millions of other customers have had no failures. You just happen to be in the failure category. It has nothing to do with China. Just look at the manufacturing issues in US. (It all came from human labor errors)
I've had generally good experiences with Apple hardware. The one Mac I owned that suffered from severe problems was plagued by its Intel CPU, so not an Apple problem at all.
A friendly advice for you - dont try to reason with pistonpilot. They seem to have issues. Let's just focus on the SSD issue.
Daaaam this looks terrible. What are the main tasks you doing? Some video editing? It's 16 or 8 gb ram?
About all those pages - I'm still gathering information, what we know for sure and what we are just "assuming". The plan is to put them into the first post as soon as possible.
To add a data-point, I have a MBP16 (32GB) bought in Dec 2019 and an M1 Mini (16GB) bought in Dec 2020.
The MBP16 has been used extensively for work for over 13 months, and the M1 Mini has been used fairly lightly as a secondary computer since purchase (and only used as a primary all-day computer for about 2 weeks since the beginning of this year).
On MBP16 smartctl data read/written were c. 163TB/170TB
smartctl results on M1 Mini:
=== START OF SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED
SMART/Health Information (NVMe Log 0x02)
Critical Warning: 0x00
Temperature: 29 Celsius
Available Spare: 100%
Available Spare Threshold: 99%
Percentage Used: 0%
Data Units Read: 41,732,486 [21.3 TB]
Data Units Written: 17,674,153 [9.04 TB]
Host Read Commands: 2,593,919,593
Host Write Commands: 89,408,074
Controller Busy Time: 0
Power Cycles: 85
Power On Hours: 439
Unsafe Shutdowns: 9
Media and Data Integrity Errors: 0
Error Information Log Entries: 0
So it does look like the M1 Mini may have slightly greater disk usage proportional to its usage time, but it is not excessive. The fact the M1 has half the RAM as the MBP could explain this.
I've now switched to using the M1 and my main desktop computer, so it will get far more regular use. I will be monitoring SMART data regularly because this is an issue of some concern. I want my computers to have at least 10 years of useful life (I tend to donate my used computers to family members after about 5 years, and don't want to give them a dud on its last legs..)
The MBP16 has been used extensively for work for over 13 months, and the M1 Mini has been used fairly lightly as a secondary computer since purchase (and only used as a primary all-day computer for about 2 weeks since the beginning of this year).
On MBP16 smartctl data read/written were c. 163TB/170TB
smartctl results on M1 Mini:
=== START OF SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED
SMART/Health Information (NVMe Log 0x02)
Critical Warning: 0x00
Temperature: 29 Celsius
Available Spare: 100%
Available Spare Threshold: 99%
Percentage Used: 0%
Data Units Read: 41,732,486 [21.3 TB]
Data Units Written: 17,674,153 [9.04 TB]
Host Read Commands: 2,593,919,593
Host Write Commands: 89,408,074
Controller Busy Time: 0
Power Cycles: 85
Power On Hours: 439
Unsafe Shutdowns: 9
Media and Data Integrity Errors: 0
Error Information Log Entries: 0
So it does look like the M1 Mini may have slightly greater disk usage proportional to its usage time, but it is not excessive. The fact the M1 has half the RAM as the MBP could explain this.
I've now switched to using the M1 and my main desktop computer, so it will get far more regular use. I will be monitoring SMART data regularly because this is an issue of some concern. I want my computers to have at least 10 years of useful life (I tend to donate my used computers to family members after about 5 years, and don't want to give them a dud on its last legs..)
I took a quick peek at my daughter's Air/M1 and there's zero swap. She's using 6 GB for programs and 6 GB for cache. I haven't checked reads/writes though I'll take a peek the next time I get to use it. She just uses it for web browsing and writing. This is a secondary system for her - she has a gaming desktop with most of her stuff.
She has a 16 GB/1 TB model. I generally get higher-spec'd systems so that they last longer.
The people with a lot of read/writes - are they doing a lot more on their M1s? In general, I try to set up my Macs so that they don't swap at all.
She has a 16 GB/1 TB model. I generally get higher-spec'd systems so that they last longer.
The people with a lot of read/writes - are they doing a lot more on their M1s? In general, I try to set up my Macs so that they don't swap at all.
It’s always good to have those inside contacts, especially with all of your replacement needs. So, if you did not know the head of USA Customer Relations your AppleCare would not have handled these latest replacements? As a self-described “damn good” consultant, do you have any recommendations as to where Apple should put its manufacturing after it immediately pulls out of China? As a self described “expert”, what leads you to believe that the manufacturing issue is with final assembly in China versus component manufacturing from Apple’s hundred of Chinese and non-Chinese sources? Or is the problem as you see it with all of them? Regarding the TBW issue, which final assembly process or componentry do you believe to be most suspect? In your professional opinion, which large scale computer manufacturer do you consider to be best of breed? I look forward to your answers. Always willing to learn from those really in the know. I can tell you have a great manufacturing perspective.
Last edited:
I'm at .9TB and I've had my Air for 5 days. 16GB/1TB model.
I've installed CC, Photoshop, Bridge, eqMac, Chrome, SteerMouse, MS365, Spotify, Bitwarden, DriveDX
I synced iCloud photos which total 12GB. I've edited 2 photos.
Email total is under 2GB.
I haven't transferred any other data to the Mac.
I've only recharged it twice.
How is almost 1TB written? My Internet data plan doesn't allow me to even use that much, so it's not from browsing, downloading, or streaming.
I have the none stop repeating error that all M1 users have, but that shouldn't write that much data.
com.apple.xpc.launchd[1] (com.apple.mdworker.shared.0B000000-0600-0000-0000-000000000000[3678]): Service exited due to SIGKILL | sent by mds[94]
Hey Apple- stop writing so much crap to the hard drive for no reason.
I've installed CC, Photoshop, Bridge, eqMac, Chrome, SteerMouse, MS365, Spotify, Bitwarden, DriveDX
I synced iCloud photos which total 12GB. I've edited 2 photos.
Email total is under 2GB.
I haven't transferred any other data to the Mac.
I've only recharged it twice.
How is almost 1TB written? My Internet data plan doesn't allow me to even use that much, so it's not from browsing, downloading, or streaming.
I have the none stop repeating error that all M1 users have, but that shouldn't write that much data.
com.apple.xpc.launchd[1] (com.apple.mdworker.shared.0B000000-0600-0000-0000-000000000000[3678]): Service exited due to SIGKILL | sent by mds[94]
Hey Apple- stop writing so much crap to the hard drive for no reason.
No video editing. Just usual apps for web development like text editor and tools like that apart from browser (Chrome and Safari), email (Spark), Docker and another few things.
MBA base with 8Gb/256Gb after 2 months
Is it normal the last error? 🤔
My iMac i9 64Gb/512Gb after one year for comparision
Code:
=== START OF SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED
SMART/Health Information (NVMe Log 0x02)
Critical Warning: 0x00
Temperature: 26 Celsius
Available Spare: 100%
Available Spare Threshold: 99%
Percentage Used: 0%
Data Units Read: 5,865,863 [3.00 TB]
Data Units Written: 4,358,300 [2.23 TB]
Host Read Commands: 83,130,495
Host Write Commands: 34,514,595
Controller Busy Time: 0
Power Cycles: 202
Power On Hours: 35
Unsafe Shutdowns: 13
Media and Data Integrity Errors: 0
Error Information Log Entries: 0
Read 1 entries from Error Information Log failed: GetLogPage failed: system=0x38, sub=0x0, code=745Is it normal the last error? 🤔
My iMac i9 64Gb/512Gb after one year for comparision
Code:
=== START OF SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED
SMART/Health Information (NVMe Log 0x02)
Critical Warning: 0x00
Temperature: 25 Celsius
Available Spare: 99%
Available Spare Threshold: 10%
Percentage Used: 0%
Data Units Read: 3,917,770 [2.00 TB]
Data Units Written: 1,121,928 [574 GB]
Host Read Commands: 7,252,370
Host Write Commands: 1,322,175
Controller Busy Time: 30
Power Cycles: 90,541
Power On Hours: 100
Unsafe Shutdowns: 46
Media and Data Integrity Errors: 5
Error Information Log Entries: 5
Warning Comp. Temperature Time: 0
Critical Comp. Temperature Time: 0
Temperature Sensor 1: 25 Celsius
Temperature Sensor 2: 26 Celsius
Error Information (NVMe Log 0x01, 16 of 64 entries)
No Errors Logged
Finally I've got my mba with 16gb of ram and 512 ssd.
So, I tested for five days, and got this numbers:
Percentage Used: 0%
Data Units Read: 659,628 [337 GB]
Data Units Written: 867,229 [444 GB]
Host Read Commands: 11,046,871
Host Write Commands: 9,258,938
Controller Busy Time: 0
Power Cycles: 82
Power On Hours: 11
Not too much.
So, I tested for five days, and got this numbers:
Percentage Used: 0%
Data Units Read: 659,628 [337 GB]
Data Units Written: 867,229 [444 GB]
Host Read Commands: 11,046,871
Host Write Commands: 9,258,938
Controller Busy Time: 0
Power Cycles: 82
Power On Hours: 11
Not too much.
Last edited:
I have no idea but looks weird only 100h in one year 🤔 I use it 2/3 hours/day.
All these stats are weirds, 35hours on the MBA that has 2 months of life? And 100hours on the iMac?! Meh...
Maybe there’s only a bug on the reading with smartcl on ARM macOS. Also because I had to move my homebrew installation in order to use/install new formulae! (Previously I had migrated from iMac to MBA ARM, but I had to remove and reinstall homebrew)
That seems low compared to mine (M1 Mini 16GB/512GB):
Code:
uptime
19:23 up 3 days, 5:41, 18 users, load averages: 2.97 3.53 3.51
iostat -Id disk0
disk0
KB/t xfrs MB
112.13 32127102 3518007.07So that's 3.5TB (read & write?) in 3 days. It's gone up by about 1TB during today's work. That's twice my SSD capacity. Even with a fair bit of swap (I've got about 15GB swap at the moment), that seems to be a large number. I'm using the same apps during the day, so I wouldn't expect a lot of swaping, but numbers from vm_stat are also very high:
Here's my Disk use from ActivityMonitor:
....and here's the output from smartctl:
So your figures of 337/443GB read/write are much lower....but you only have 11 hours on power-on time. Is that correct? Presumably SMART data is total disk usage for the life of the machine, and not reset on reboot.
I'm now starting to get a bit worried. If I'm writing 500-1000GB to the SSD every day in "normal" OS usage (not copying huge files around) then how long would the SSD last? What is the TBW for these drives?
Attachments
I think not. There are over 30hr on Activity Monitor. I think sysmontools shows wrong numbers
I disabled Siri 20 days ago on the MBA and I have never enabled it on the iMac. Doesn’t look the cause.
Is the kernel task process that is writing a lot but my swap space remains low (1/2gb max)