You are not logged in.
I'm currently using Arch, but ive had this issue on every single linux distro I've tried on this computer. I'm using an nvme ssd, and whenever I stress it (like by moving large files between it and my other nvme ssd), my computer becomes unresponsive and practically unusable until the file transfer is done. It's almost like nothing else can properly access the computer's storage IO until the operation is done. This happens regardless of how I try to move or copy the files, whether through a command like mv or rsync or a file manager such as nemo or dolphin. I'm doing a large file transfer on the computer I'm typing this on right now which is stressing the drive, and if it weren't for the fact that my browser window were already open it would've taken like 30 minutes to get to this site if it loaded at all. I'm on kde plasma and plasmashell just completely crashes and often refuses to start up again until theres no more file transfers and its the same for a lot of programs. Even programs that are open will freeze and crash constantly. Everything is working perfectly fine with no issues until I try a fast file transfer. That makes everything instantly fall apart. Is there a way to remedy this behavior? I've not managed to find anything about this in my searching.
Also, if I'm booted off of a slower drive such as an HDD or a usb attached drive, the problem is 500x worse. Everything will just freeze randomly regardless of whether I'm moving files, again on every distro I've tried. I don't think this behavior existss to this extent on other computers I own, it just seems to be this one and I can't figure out why.
Last edited by queenbiscuit311 (2023-07-26 18:11:01)
Offline
Post some info about this ssd (manufacturer, model), post output of commands:
smartctl -A /dev/device_name like /dev/nvme0
hdparm -iI /dev/device_name
nvme-smart-log /dev/device_name
nvme-list
nvme-list-ctrl /dev/device_name
nvme device-self-test /dev/nvme0 -n 1 -s 1
journalctl -b (full from bot where you experience slowdown)
Offline
I'm not sure which boot was the last where I had the issue, but I'll make sure to send everything next time i encounter the issue. I'll see if I can recreate it by just moving random stuff.
Last edited by queenbiscuit311 (2023-07-26 20:13:49)
Offline
Post some info about this ssd (manufacturer, model), post output of commands:
My boot drive is a 1TB Kingston NVMe drive. According to the gnome disk utility, its model is "KINGSTON OM8PCP31024F-AI1 (ECFK52.8)". My Arch installation is on a BTRFS formatted partition, but the problem exists to the same extent on EXT4.
I don't feel like using an infinite amount of code tags, so I'll link to pastebins with the command outputs.
smartctl -A /dev/nvme0n1: https://pastebin.com/LwyGuKbg
hdparm fails with error "HDIO_GET_IDENTITY failed: Inappropriate ioctl for device", probably because I'm not on an ATA drive.
nvme smart-log /dev/nvme0n1: https://pastebin.com/CCUK0QRi
nvme list: https://pastebin.com/mK7rE7c8
nvme list-ctrl /dev/nvme0n1:
num of ctrls present: 2
[ 0]:0x1
[ 1]:0x2
nvme device-self-test /dev/nvme0n1 -n 1 -s 1:
Short Device self-test started
Trying to export my journalctl the first time decided to freeze everything and crash my computer, and then trying to export it after rebooting, its made a 1.4 GiB text file. I'm not entirely sure what to do with this monstrosity. Most of it seems to be nvidia drivers spitting out the same warning message ad nauseam.
Last edited by queenbiscuit311 (2023-07-26 21:02:48)
Offline
https://wiki.archlinux.org/title/System … ing_output
Can you reproduce this w/
- openbox
- dd if=/path/to/file -of path/to/copy -oflags=direct
- … out zswap, https://wiki.archlinux.org/title/Zswap#Toggling_zswap
Esp. in case of the latter two see https://lonesysadmin.net/2013/12/22/bet … rty_ratio/ and limit the cache
Do the other system(s) have
1. as much RAM
2. an nvidia GPU?
Offline
Your nvme ssd temperature is high (almost 70 degree, it may be almost at operational limit), your drive may overheating and throttle. Do you have radiator for it? Is it hot where your computer is? You may want to check temperatures of other components: CPU, GPU, RAM.
Offline
Is the drive encrypted? There's a kernel bug that causes a deadlock and has been getting worse since 5.x.x.
I've been using priorities and setting the CPU scheduler to performance to try and mitigate it.
alias mv='/bin/nice -n +20 /bin/mv'Offline