Swap issue...but swap disabled...? So confused!

[darkstrike]

Hi all - having an issue with mom's laptop. It's a Toshiba running Mint 20.2 MATE 64-bit with a Core i7 CPU and 8GB of RAM and a 500GB SSD. It's boggling my mind!

Basically, fairly often the whole system will freeze and the HDD activity light will just start flashing CONSTANTLY and start running molasses slow....as in swiping the mouse from one side of the screen to the other is like a slide show where the mouse is moving a few pixels every 10 seconds...

She's an older lady in her 70s, so she doesn't do anything fancy on it other than email and youtube and posting things for sale on Kijiji and Facebook Marketplace (the only thing she uses Facebook for).

The only time I've ever seen that behaviour on my own PCs with Mint was when RAM got used up and the machine started HEAVILY dipping into swap - it even did this when I had swappiness set to '1', so I have swap completely disabled on my own laptop and haven't really had the problem since (I have 12GB of RAM in my own laptop).

Now, 8GB should be more than enough for most normal stuff on her laptop, so I thought I'd try disabling the swap on her system as well to see if things would improve...used the swapoff /a command in terminal to stop it, deleted the swap reference in the fstab file and deleted the swap partition and absorbed it into the main partition.

BUT, fairly often it still behaves like it's suddenly run out of memory and started using swap as described above! I'm at a loss as to what is going on or how to fix it...

The most for programs she has running at a time are maybe Discord and Firefox, and sometimes LibreOffice Writer at the same time as well? Discord and Libre take up relatively little memory and I know that browsers can be memory hogs, but she only ever has at most, like 6-7 tabs open (it's usually only 4-5), which is barely anything (I have 27 tabs open on my 12GB RAM also Core i7 laptop also running Mint 20.2 MATE right now and still have 30% of my RAM free...)

Does Mint have an 'emergency' swap that somehow turns itself on even if swap is fully disabled...?

And yes, I know you'll all say 'why are you disabling swap', but my point is that it is somehow coming on unnecessarily when nowhere near all her RAM is used up even with swap DISABLED and very few programs running and leading to the entire system having to be forcibly restarted...

Any folks have any ideas?

The only OTHER thing I can think of is that the SATA port on her laptop is broken, so the SSD is installed in one of those Optical-to-SATA adapters in the CD drive bay. I've even tried using two different optical-to-SATA adapters and that didn't help either. The SSD is a brand new Samsung EVO 870 that is only 5-6 months old!

Thanks for any suggestions - I'm at my wit's end!

[elixxx]

hi, please share output of following terminal commands during system freezing incident you described.

Code: Select all

cat /proc/sys/vm/swappiness
htop

[Pjotr]

Also, please generate an overview of the system like this:

- Launch a terminal window (this is how to launch a terminal window);

- Make the terminal window full screen, to avoid chopped lines;

- Copy/paste this command into the terminal:

Code: Select all

inxi -Fxpmrz

(if you type: the letter F is a capital letter, and don't omit the space after inxi!)

Press Enter.

Copy/paste the output in your next message.

[jjp2145-oldtimer]

Somebody asked for the result of the inxi command while I was typing.

Code: Select all

free -h

can be used to show how much memory and swap is being used. If swap is not in use, it will not be listed.

[darkstrike]

hi, please share output of following terminal commands during system freezing incident you described.

Code: Select all

cat /proc/sys/vm/swappiness
htop

I can't...the system freezes and the mouse moves so slowly it may as well be fully frozen. I can't get to the terminal to run any commands when it happens.

[zcot]

Just a suggestion, because I saw what you posted in that other thread, that you're not quite catching the nuances.

You should just recreate the file, make it 8GB this time, how about that, instead of the default 1GB, and turn it back on, and don't mess with the swappiness value, put it back to 60, that's not a junk drive that's going to get killed by usage. If she needs it, then the system is broken without it.

You mention "swap is somehow coming back on", but that's not likely. What can be happening, and probably is in this case, is something completely non-related to swap in the first place.

Disabling swap is not a functional method for keeping away a slow-down, and especially if it's actually related to ram exhaustion.

Also, with an i7 and SSD, and getting into, say, 80% of ram usage and fully committed into swap usage, it should barely even be consequential to the user experience.

But what about that adapter?

Get the system info before it gets bogged down.

On a side note, maybe I saw you mention using a system monitor or something watching your ram, -your laptop, with swap disabled. You should do a test and watch that 80% but just keep loading it up a bit more until you get it to 100% and see what happens. I'm just saying it because if you are not actively monitoring, which I would say your mom is not, aside from whatever issue is actually coming up here, then that's the perfect candidate to have swap in place.

[darkstrike]

Have to head to work, but I'll try and get the info you requested later today from her laptop when I get home!

[darkstrike]

hi, please share output of following terminal commands during system freezing incident you described.

Code: Select all

cat /proc/sys/vm/swappiness
htop

Results of the swappiness command....again, I can't run these when it's frozen as it's effectively incapable of doing things until restart it, but these command results are from when it WAS running fine this evening.

Code: Select all

1

Because I had set it to this to prevent swap from cutting in.

Htop:

[darkstrike]

Also, please generate an overview of the system like this:

- Launch a terminal window (this is how to launch a terminal window);

- Make the terminal window full screen, to avoid chopped lines;

- Copy/paste this command into the terminal:

Code: Select all

inxi -Fxpmrz

(if you type: the letter F is a capital letter, and don't omit the space after inxi!)

Press Enter.

Copy/paste the output in your next message.

System info results:

Code: Select all

donna@donnalaptop:~$ inxi -Fxpmrz
System:
  Kernel: 5.4.0-86-generic x86_64 bits: 64 compiler: gcc v: 9.3.0 
  Desktop: MATE 1.24.0 Distro: Linux Mint 20.2 Uma base: Ubuntu 20.04 focal 
Machine:
  Type: Laptop System: TOSHIBA product: Satellite L650 v: PSK1EC-0CW03D 
  serial: <filter> 
  Mobo: TOSHIBA model: Portable PC serial: <filter> BIOS: INSYDE v: 1.90 
  date: 11/11/2010 
Battery:
  ID-1: BAT0 charge: 0.0 Wh condition: 3.2/48.6 Wh (7%) model: PA381 
  status: Unknown 
Memory:
  RAM: total: 7.58 GiB used: 2.33 GiB (30.8%) 
  RAM Report: 
  permissions: Unable to run dmidecode. Root privileges required. 
CPU:
  Topology: Dual Core model: Intel Core i5 M 480 bits: 64 type: MT MCP 
  arch: Nehalem rev: 5 L2 cache: 3072 KiB 
  flags: lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx bogomips: 21332 
  Speed: 2933 MHz min/max: 1199/2667 MHz Core speeds (MHz): 1: 2933 2: 2933 
  3: 2933 4: 2933 
Graphics:
  Device-1: Intel Core Processor Integrated Graphics vendor: Toshiba 
  driver: i915 v: kernel bus ID: 00:02.0 
  Display: x11 server: X.Org 1.20.11 driver: modesetting 
  unloaded: fbdev,vesa resolution: 1366x768~60Hz 
  OpenGL: renderer: Mesa DRI Intel HD Graphics (ILK) v: 2.1 Mesa 21.0.3 
  direct render: Yes 
Audio:
  Device-1: Intel 5 Series/3400 Series High Definition Audio vendor: Toshiba 
  driver: snd_hda_intel v: kernel bus ID: 00:1b.0 
  Sound Server: ALSA v: k5.4.0-86-generic 
Network:
  Device-1: Qualcomm Atheros AR8152 v1.1 Fast Ethernet vendor: Toshiba 
  driver: atl1c v: 1.0.1.1-NAPI port: 3000 bus ID: 01:00.0 
  IF: enp1s0 state: down mac: <filter> 
  Device-2: Qualcomm Atheros AR928X Wireless Network Adapter vendor: Lite-On 
  driver: ath9k v: kernel port: 3000 bus ID: 02:00.0 
  IF: wlp2s0 state: up mac: <filter> 
Drives:
  Local Storage: total: 232.89 GiB used: 68.62 GiB (29.5%) 
  ID-1: /dev/sda vendor: Samsung model: SSD 870 EVO 250GB size: 232.89 GiB 
Partition:
  ID-1: / size: 227.73 GiB used: 68.62 GiB (30.1%) fs: ext4 dev: /dev/sda5 
  ID-2: /boot/efi size: 511.0 MiB used: 4 KiB (0.0%) fs: vfat dev: /dev/sda1 
  ID-3: /run/timeshift/backup raw size: 232.38 GiB 
  size: <superuser/root required> used: <superuser/root required> fs: ext4 
  dev: /dev/sda5 
Sensors:
  System Temperatures: cpu: 72.0 C mobo: N/A 
  Fan Speeds (RPM): N/A 
Repos:
  No active apt repos in: /etc/apt/sources.list 
  Active apt repos in: /etc/apt/sources.list.d/google-chrome.list 
  1: deb [arch=amd64] http://dl.google.com/linux/chrome/deb/ stable main
  Active apt repos in: /etc/apt/sources.list.d/official-package-repositories.list 
  1: deb https://muug.ca/mirror/linuxmint/packages uma main upstream import backport
  2: deb http://mirror.its.dal.ca/ubuntu focal main restricted universe multiverse
  3: deb http://mirror.its.dal.ca/ubuntu focal-updates main restricted universe multiverse
  4: deb http://mirror.its.dal.ca/ubuntu focal-backports main restricted universe multiverse
  5: deb http://security.ubuntu.com/ubuntu/ focal-security main restricted universe multiverse
  6: deb http://archive.canonical.com/ubuntu/ focal partner
  No active apt repos in: /etc/apt/sources.list.d/skype-stable.list 
  Active apt repos in: /etc/apt/sources.list.d/teamviewer.list 
  1: deb https://linux.teamviewer.com/deb stable main
  Active apt repos in: /etc/apt/sources.list.d/tsbarnes-indicator-keylock-focal.list 
  1: deb http://ppa.launchpad.net/tsbarnes/indicator-keylock/ubuntu focal main
Info:
  Processes: 232 Uptime: 1d 10h 42m Init: systemd runlevel: 5 Compilers: 
  gcc: 9.3.0 Shell: bash v: 5.0.17 inxi: 3.0.38 

[darkstrike]

Somebody asked for the result of the inxi command while I was typing.

Code: Select all

free -h

can be used to show how much memory and swap is being used. If swap is not in use, it will not be listed.

Currently seems to show no swap:

Code: Select all

donna@donnalaptop:~$ free -h
              total        used        free      shared  buff/cache   available
Mem:          7.6Gi       2.1Gi       3.7Gi       303Mi       1.7Gi       4.8Gi
Swap:            0B          0B          0B

[darkstrike]

Just a suggestion, because I saw what you posted in that other thread, that you're not quite catching the nuances.

You should just recreate the file, make it 8GB this time, how about that, instead of the default 1GB, and turn it back on, and don't mess with the swappiness value, put it back to 60, that's not a junk drive that's going to get killed by usage. If she needs it, then the system is broken without it.

You mention "swap is somehow coming back on", but that's not likely. What can be happening, and probably is in this case, is something completely non-related to swap in the first place.

Disabling swap is not a functional method for keeping away a slow-down, and especially if it's actually related to ram exhaustion.

Also, with an i7 and SSD, and getting into, say, 80% of ram usage and fully committed into swap usage, it should barely even be consequential to the user experience.

But what about that adapter?

Get the system info before it gets bogged down.

On a side note, maybe I saw you mention using a system monitor or something watching your ram, -your laptop, with swap disabled. You should do a test and watch that 80% but just keep loading it up a bit more until you get it to 100% and see what happens. I'm just saying it because if you are not actively monitoring, which I would say your mom is not, aside from whatever issue is actually coming up here, then that's the perfect candidate to have swap in place.

On my own system, I have a pretty sweet Conky setup on my desktop with a usage bar for RAM, so I tend to close and reopen Firefox when I notice RAM use getting high (unlike mom though, who usually has no more than 6-7 tabs open, I often have like....25-30, but I also have 12GB of RAM where she has 8GB, haha! I do also use Firefox though, as a browser.)

Mom has no monitor set up for RAM use though...

The reason I thought it is swap is because it will be fine one minute, then I just see the HDD activity light come on and flicker constantly so much so that it looks almost like it's solidly on and then the system slows to a CRAWL.....it's not TECHNICALLY frozen, as I can still move the mouse, but the cursor moves VERY slowly and has HORRIBLE latency....until I restart the machine and it goes back to normal! As for the adapter....I've tried two different ones now...slightly different models from different brands as the adapter was originally what I thought was the issue, but the machine runs GREAT when it's working! I'm typing this from it now! It's just that it has these random hang-ups that she has lost a lot of work due to in programs with no propensity for saving as you go!

I posted the sytem info for this laptop a few posts back....you can see it there!

[motoryzen]

ID-3: /run/timeshift/backup raw size: 232.38 GiB

This has me curious. That samsung evo ssd is too young a drive to be having problems save for it ever reaching to having around 20% remaining capacity or less...left, but I don't think that is the case.

1. Do you not direct your Timeshift backups to a separate physical storage drive AND is it currently directed to target that same only samsung evo ssd? If yes, I recommend you change that ( and yes..I agree with the creator/creators/dev maintainers of Timeshift would NOT have this be the default behavior OOTB of installing Linux Mint...makes no sense to me at all).

2. Check and see how many timeshift snapshots you have. If you have ...I'd say more than 5, consider deleting the oldest up until the 5th or 4th one and see if that helps. Yes I see that only around 30 something percent capacity of that drive is consumed..but still. IT's worth looking at.

3. Given you're on a laptop, does this problem only happen when you're on battery only? And if so only near the last 7% battery charge remaining? Might not be coherent, but I like to rule out everything possible.

Basically, fairly often the whole system will freeze and the HDD activity light will just start flashing CONSTANTLY and start running molasses slow....as in swiping the mouse from one side of the screen to the other is like a slide show where the mouse is moving a few pixels every 10 seconds..

How often does it happen...everyday? I'm wondering if timeshift on that laptop is set to do daily backups /the same days this problem occurs.

[zcot]

Code: Select all

  Speed: 2933 MHz min/max: 1199/2667 MHz Core speeds (MHz): 1: 2933 2: 2933 
  3: 2933 4: 2933 

That cpu is really committed to the job! Can we call that all 110%?

I can't say why, but I've seen that on a number of occasions, sometimes it's the particular kernel and there's some regression that's causing an issue. That's a screaming cpu, and as soon as more commands get thrown at it, at some point there's a bottleneck scenario.

Is that just running high because of that kernel vs that hardware setup? maybe.

Is it due to some background service that isn't playing nice? maybe.

Is the tsbarnes indicator-keylock working well? that's the only odd thing that I can see from the output. I do see the dev daily ppa doesn't have any entries from the focal base. It was packaged for focal and probably appears to function fine. It would be an easy test anyway to remove that.

Otherwise it's digging in the syslog. /var/log/syslog maybe you give a quick scroll and see something obvious.

[absque fenestris]

Sensors:
System Temperatures: cpu: 72.0 C mobo: N/A
Fan Speeds (RPM): N/A

That's not exactly cool either - would at least also be a circumstance to check. Case and fan cleaning, for example?
Then there are also TeamViewer and Skype in the game. Two candidates that would have to be checked critically.

[darkstrike]

ID-3: /run/timeshift/backup raw size: 232.38 GiB

This has me curious. That samsung evo ssd is too young a drive to be having problems save for it ever reaching to having around 20% remaining capacity or less...left, but I don't think that is the case.

1. Do you not direct your Timeshift backups to a separate physical storage drive AND is it currently directed to target that same only samsung evo ssd? If yes, I recommend you change that ( and yes..I agree with the creator/creators/dev maintainers of Timeshift would NOT have this be the default behavior OOTB of installing Linux Mint...makes no sense to me at all).

2. Check and see how many timeshift snapshots you have. If you have ...I'd say more than 5, consider deleting the oldest up until the 5th or 4th one and see if that helps. Yes I see that only around 30 something percent capacity of that drive is consumed..but still. IT's worth looking at.

3. Given you're on a laptop, does this problem only happen when you're on battery only? And if so only near the last 7% battery charge remaining? Might not be coherent, but I like to rule out everything possible.

Basically, fairly often the whole system will freeze and the HDD activity light will just start flashing CONSTANTLY and start running molasses slow....as in swiping the mouse from one side of the screen to the other is like a slide show where the mouse is moving a few pixels every 10 seconds..

How often does it happen...everyday? I'm wondering if timeshift on that laptop is set to do daily backups /the same days this problem occurs.

Issue happens once every day or two...sometimes 3 days.

Never on battery - the battery is shot, so it's ALWAYS plugged in (when unplugged the battery lasts maybe 10 mins?)

Opening TimeShift only shows 3 restore points - it is only set to do one weekly backup on Fridays...? Timeshift saves to the same drive the OS/Data is on because it's a laptop, so no other internal drive and no external drive connected. The SSD is installed in one of those optical-to-SATA adapter trays because the actual SATA hard drive port on the laptop's motherboard is physically broken. Each TimeShift restore point is about 9GB I think...?

I've checked the Samsung SSD's drive health using GSmartControl....it's still showing something like 98 or 99% lifetime remaining.

[lofi]

[Swap issue, but swap disabled]
Basically, fairly often the whole system will freeze and the HDD activity light will just start flashing CONSTANTLY and start running molasses slow....as in swiping the mouse from one side of the screen to the other is like a slide show where the mouse is moving a few pixels every 10 seconds...
(...)
The only time I've ever seen that behaviour on my own PCs with Mint was when RAM got used up and the machine started HEAVILY dipping into swap (...).

This is highly evocative of,
- System freeze when memory is put on SWAP in Linux >4.10.x (also when no swap) https://bugs.launchpad.net/ubuntu/+sour ... ug/1833281 (good starting point)
- https://bugzilla.kernel.org/show_bug.cgi?id=199763
- https://bugzilla.kernel.org/show_bug.cgi?id=196729 (even if the original report is about swap, people with disabled swap are affected)
- linked to this legendary closed bug https://bugzilla.kernel.org/show_bug.cgi?id=12309 from around 2007
- funny page in Russian about bug 12309 http://lurkmore.to/12309

I'm not sure it's that, but it's highly evocative. In this bug, swap can trigger the freezes, but is not directly responsible. It would be a deeper I/O writes problem. I don't know how to summarize these links, as to why it also happens when swap is disabled, but in the same circumstances (stress on RAM). But it's mentioned in the links.

Fixes mentioned are,
- raise /proc/sys/vm/vfs_cache_pressure to 6000
- bump vm.min_free_kbytes way up to 393216

[lofi]

I found a source for why kswapd (the swapping daemon) could be responsible even when swap is disabled :

Swap space is only used for data that is not backed by any other file. Data that is mapped from other files on disk ( such as executable programs ) is still swapped to their respective files even if you don't have a swap device.

https://askubuntu.com/questions/432809/ ... 827#432827

[Larry78723]

Code: Select all

  Speed: 2933 MHz min/max: 1199/2667 MHz Core speeds (MHz): 1: 2933 2: 2933 
  3: 2933 4: 2933 

That cpu is really committed to the job! Can we call that all 110%?

I can't say why, but I've seen that on a number of occasions, sometimes it's the particular kernel and there's some regression that's causing an issue. That's a screaming cpu, and as soon as more commands get thrown at it, at some point there's a bottleneck scenario.

Is that just running high because of that kernel vs that hardware setup? maybe.

Is it due to some background service that isn't playing nice? maybe.

Is the tsbarnes indicator-keylock working well? that's the only odd thing that I can see from the output. I do see the dev daily ppa doesn't have any entries from the focal base. It was packaged for focal and probably appears to function fine. It would be an easy test anyway to remove that.

Otherwise it's digging in the syslog. /var/log/syslog maybe you give a quick scroll and see something obvious.

I would check to see if Team Viewer and Skype are enabled in Startup Apps. If so, try disabling them one at a time and see if either one is misbehaving. If that's not the situation, I'd boot into an earlier kernel and see if the condition improves.

[darkstrike]

Code: Select all

  Speed: 2933 MHz min/max: 1199/2667 MHz Core speeds (MHz): 1: 2933 2: 2933 
  3: 2933 4: 2933 

That cpu is really committed to the job! Can we call that all 110%?

I can't say why, but I've seen that on a number of occasions, sometimes it's the particular kernel and there's some regression that's causing an issue. That's a screaming cpu, and as soon as more commands get thrown at it, at some point there's a bottleneck scenario.

Is that just running high because of that kernel vs that hardware setup? maybe.

Is it due to some background service that isn't playing nice? maybe.

Is the tsbarnes indicator-keylock working well? that's the only odd thing that I can see from the output. I do see the dev daily ppa doesn't have any entries from the focal base. It was packaged for focal and probably appears to function fine. It would be an easy test anyway to remove that.

Otherwise it's digging in the syslog. /var/log/syslog maybe you give a quick scroll and see something obvious.

I would check to see if Team Viewer and Skype are enabled in Startup Apps. If so, try disabling them one at a time and see if either one is misbehaving. If that's not the situation, I'd boot into an earlier kernel and see if the condition improves.

TeamViewer IS enabled as I use it to remotely access mom's laptop from afar when I am out and about and she has trouble. Skype is NOT set to run on Startup.

@zcot - The keyboard status indicator works great - her laptop has an LED indicator light in the capslock key, but the indicator LED in the keyboard for Numlock is burned out, so I use that app to let her know when she has Numlock on.

[darkstrike]

I found a source for why kswapd (the swapping daemon) could be responsible even when swap is disabled :

Swap space is only used for data that is not backed by any other file. Data that is mapped from other files on disk ( such as executable programs ) is still swapped to their respective files even if you don't have a swap device.

https://askubuntu.com/questions/432809/ ... 827#432827

Wonderful....so does that mean if I put swap BACK it would help the situation...? Not sure I understand? I had lowered the swap use originally way back to help save on wear on the SSD.