Solus not utilizing full RAM available, only 16 of 64

zinger565

Ran into an issue that's been occurring over the last month or so. Solus only seems to be using about 16Gb of ram before filling Swap, and when that fills up, it starts to kill applications until it kills itself.

I have 4 sticks of 16Gb ram installed, 2 matched pairs, installed correctly. Motherboard BIOS is able to see it all, and Solus (via system monitor and the Budgie "about" section) sees it just fine. The hardware itself has been installed and unchanged for 10 months, and Solus had no problem with it until recently (maybe October?).

And no, 16Gb isn't "enough", this is causing serious workflow issues for photo and video editing (Darktable, TopazAI, and KdenLive), to the point the system will crash mid-export.

What other information do I need to provide and what are steps to correct?

Budgie version is 10.9.2
Solus 4.6 Convergence
Kernel Linux 6.11.10-310.current

brent

Axios If it gives you trouble again take pencil eraser and clean the contacts lightly.

that's brilliant. that takes any debris or coating right off so it can contact the slot better. I would've tried that if I knew! In hindsight I wonder if it would be also smart to blow the narrow bay slots off with compressed air? Or overkill?

either way I hope OP tries some of these ideas.

zinger565 Solus only seems to be using about 16Gb of ram before filling Swap, and when that fills up, it starts to kill applications until it kills itself.

is that way its supposed to work I wonder?

zinger565 have 4 sticks of 16Gb ram installed, 2 matched pairs, installed correctly.

that's what I thought to when I put my rig together recently 🙂

zinger565 The hardware itself has been installed and unchanged for 10 months, and Solus had no problem with it until recently (maybe October?

do buzz's memtest along with the commands. you will get a clearer picture.

BuzzPCSOS

First test your RAM. The bios seeing it does not prove that there is not a stick failing that could cause your problem.
Download a bootable Linux iso (64 bits) from Memtest+ to get a test which is independent of your operating system.

brent

I find this complex. bad stick? bad bank? who knows.
cat /proc/meminfo will show you what the computer sees, here is mine.
MemTotal: 12033284 kB
MemFree: 4699736 kB
MemAvailable: 7842824 kB

I have 16GB so something is wrong here. It sees 12....

$ sudo lshw -class memory
*-firmware
description: BIOS
vendor: LENOVO
physical id: 0
version: FBKTE0AUS
date: 12/23/2021
size: 64KiB
capacity: 6656KiB
capabilities: pci upgrade shadowing cdboot bootselect socketedrom edd int13floppy1200 int13floppy720 int13floppy2880 int5printscreen int9keyboard int14serial int17printer acpi usb biosbootspecification uefi
*-cache:0
description: L2 cache
physical id: 3e
slot: CPU Internal L2
size: 1MiB
capacity: 1MiB
capabilities: internal write-back unified
configuration: level=2
*-cache:1
description: L1 cache
physical id: 3f
slot: CPU Internal L1
size: 256KiB
capacity: 256KiB
capabilities: internal write-back
configuration: level=1
*-cache:2
description: L3 cache
physical id: 40
slot: CPU Internal L3
size: 6MiB
capacity: 6MiB
capabilities: internal write-back unified
configuration: level=3
*-memory
description: System Memory
physical id: 41
slot: System board or motherboard
size: 12GiB
*-bank:0
description: DIMM DDR3 Synchronous 1600 MHz (0.6 ns)
product: M378B5173EB0-YK0
vendor: Samsung
physical id: 0
serial: 21D1CB21
slot: ChannelA-DIMM0
size: 4GiB
width: 64 bits
clock: 1600MHz (0.6ns)
*-bank:1
description: DIMM DDR3 Synchronous 1600 MHz (0.6 ns)
product: M378B5273EB0-CK0
vendor: Samsung
physical id: 1
serial: 9598FB7A
slot: ChannelA-DIMM1
size: 4GiB
width: 64 bits
clock: 1600MHz (0.6ns)
*-bank:2
description: DIMM DDR3 Synchronous 1600 MHz (0.6 ns)
product: M378B5173EB0-YK0
vendor: Samsung
physical id: 2
serial: 21D1C39A
slot: ChannelB-DIMM0
size: 4GiB
width: 64 bits
clock: 1600MHz (0.6ns)
*-bank:3
description: DIMM [empty]
product: [Empty]
vendor: [Empty]
physical id: 3
serial: [Empty]
slot: ChannelB-DIMM1

I am shocked to find out for myself that bank 3 is "EMPTY" since I have another 4GB in there. I'm finding out a he** of a lot more about my own situation than yours, but you can see if you have a match between the cat command and the lshw command..

brent

brent
cat /proc/meminfo now show 16 GB and so does LSHW...
@BuzzPCSOS can confirm the BIOS reading you mentioned.

SOLUTION FOR ME: Taking out the Bay 3 stick and re-seating it. It did not seem off or uneven or loose in any way but evidently it was not seated exactly where the MOBO like it..
...you could try physically re-seating them. It's more likely that than anything imo.
2 cents and good luck.

zinger565

brent

Here's mine:

MemTotal: 64955308 kB
MemFree: 14599964 kB
MemAvailable: 58451888 kB
Buffers: 534552 kB
Cached: 43007152 kB

*-firmware
description: BIOS
vendor: American Megatrends International, LLC.
physical id: 0
version: F10
date: 07/10/2023
size: 64KiB
capacity: 32MiB
capabilities: pci upgrade shadowing cdboot bootselect socketedrom edd int13floppynec int13floppytoshiba int13floppy360 int13floppy1200 int13floppy720 int13floppy2880 int5printscreen int14serial int17printer int10video acpi usb biosbootspecification uefi
*-cache:0
description: L1 cache
physical id: a
slot: L1 - Cache
size: 1MiB
capacity: 1MiB
clock: 1GHz (1.0ns)
capabilities: pipeline-burst internal write-back unified
configuration: level=1
*-cache:1
description: L2 cache
physical id: b
slot: L2 - Cache
size: 16MiB
capacity: 16MiB
clock: 1GHz (1.0ns)
capabilities: pipeline-burst internal write-back unified
configuration: level=2
*-cache:2
description: L3 cache
physical id: c
slot: L3 - Cache
size: 64MiB
capacity: 64MiB
clock: 1GHz (1.0ns)
capabilities: pipeline-burst internal write-back unified
configuration: level=3
*-memory
description: System Memory
physical id: f
slot: System board or motherboard
size: 64GiB
*-bank:0
description: DIMM Synchronous Unbuffered (Unregistered) 5600 MHz (0.2 ns)
product: CP16G56C46U5.C8D
vendor: Unknown
physical id: 0
serial: E9142E7D
slot: DIMM 0
size: 16GiB
width: 64 bits
clock: 1305MHz (0.8ns)
*-bank:1
description: DIMM Synchronous Unbuffered (Unregistered) 5600 MHz (0.2 ns)
product: CP16G56C46U5.C8D
vendor: Unknown
physical id: 1
serial: E9142F16
slot: DIMM 1
size: 16GiB
width: 64 bits
clock: 1305MHz (0.8ns)
*-bank:2
description: DIMM Synchronous Unbuffered (Unregistered) 5600 MHz (0.2 ns)
product: CP16G56C46U5.C8D
vendor: Unknown
physical id: 2
serial: E9142E91
slot: DIMM 0
size: 16GiB
width: 64 bits
clock: 1305MHz (0.8ns)
*-bank:3
description: DIMM Synchronous Unbuffered (Unregistered) 5600 MHz (0.2 ns)
product: CP16G56C46U5.C8D
vendor: Unknown
physical id: 3
serial: E9143092
slot: DIMM 1
size: 16GiB
width: 64 bits
clock: 1305MHz (0.8ns)

Whatever recent update was pushed seems to have helped. I've seen usage up to 50% now without SWAP filling up.
Is the "cache" shown on SystemMonitor also "in memory"? As in, if the active plus cached exceeds memory total, then SWAP starts to fill up? Is there any way to dump cache?

Edit to add: Will run the Memtest+ tonight

Axios

brent If it gives you trouble again take pencil eraser and clean the contacts lightly.
I have a dell that does that if I leave it off to long sometimes I got reseat the memory but I dont use that
computer anymore.

brent

zinger565 Whatever recent update was pushed seems to have helped. I've seen usage up to 50% now without SWAP filling up.

so it sees 64GB but you can't use it all...

Is the "cache" shown on SystemMonitor also "in memory"? As in, if the active plus cached exceeds memory total, then SWAP starts to fill up? Is there any way to dump cache?"

excellent questions. i don't know.

Staudey

zinger565 MemTotal: 64955308 kB
MemFree: 14599964 kB
MemAvailable: 58451888 kB
Buffers: 534552 kB
Cached: 43007152 kB

Well, looks like all of your RAM is getting used then. (well 14 GB free in this particular case, but that would presumably fill up too). It should generally drop the cache before swapping though.

https://www.linuxatemyram.com/

brent

maybe a dev will step in I hope.

Solus will use 32GB of your 64GB before spilling into Swap partition.

This hinders your big app usage.

figure out why 32GB is the magic number to activate Swap--instead of the full 64GB--and that's the end of the mystery.
seems like a lot factors as to why but I don't know what any of them are.

zinger565

MEMTEST+ came back with a pass.

I guess it's a "cache" issue.

BuzzPCSOS

zinger565 Can't be much help here as my own RAM tops out at 8GB. So a hunch rather than anything else.
Do you still find that there is a RAM cap if you boot and run from the Solus 4.6 iso?
Could the default Zram settings be having an effect here?
How was the original install set up? No swap, swap - no hibernate, swap with hibernate, swap from a file?
Like I say, nothing I can test here but certainly something that might give us food for thought.