Severe lag during disk IO

Inxi -Fx output is here on pastebin.

Context: when I am writing large files (e.g., tar -xvzf somelargefile, installing a game, cp largeFile), my desktop starts lagging. I have used Ubuntu 18.04 + KDE for the past 6 months with no trouble. By lagging, I mean (for example) when instantiating a new terminal window, the contents do not render. If I drag a Chrome window, it does not respond. These lags persist for the most part (during any file transfer), although there are sometimes windows of laglessness.

I have followed similar threads such as manjaro kde freezes when copying large files amongst others to no avail. Below is a list of related notes including solutions from other threads.

Other related notes:

  1. I noticed this lag occurs when my system load reads: Cache Dirty 1 GiB, Writeback at 500MiB. (Rough numbers)
  2. I have tried changing my IO schedulers to mq-deadline, bfq, none (I rebooted each time, grepped the scheduler to double check it has changed) to no avail.
  3. The swap is never used (0/32GB). I was nowhere near using up too much RAM (swappiness at 60; have 16GB RAM).
  4. This occurs whether I use the terminal or dolphin to copy files, or terminal to tar some files of size 10GB for example.
  5. This occurs with kernel versions 4.14.183-1, 4.19.126-1, 5.4.44-1 (All of which are labelled LTS), and the latest 5.7.9-1.
  6. I turned off the compositor to no avail.
  7. Changed compositor to OpenGL 3.1, OpenGL 2, XRender to no avail.
  8. my drive has the root and home partitions fully encrypted.
  9. this happens from a fresh install (I have yet to confirm whether it lags prior pacman -Syu after the fresh install).
  10. Baloo is disabled.
  11. tlp is completely removed from my computer (its a desktop).

Any ideas what could be going on? Please let me know if there is any additional information I could provide to help this debugging process.

I had success working around this type of issue by changing my schedulers. However, in the end I determined it was an old hard drive that was causing the problem in the first place. I don't know if it was because the drive was failing or it was because it contained too many illegal file names. I just know that my I/O issues and freezes went away after removing the drive. The drive passed all smartmon tests, so that is not a good way of eliminating a drive as a possibility.

If you have tlp installed I would suggest masking it as a troubleshooting step (just to see if it's involved).

Hi tbg, thanks for the suggestions!

My drive is an nvme SSD which worked fine in Ubuntu for whatever reason, so it's probably unrelated to the drive...(?) how should I go about confirming this?

i do have TLP installed, why is it used on a desktop? I thought it was for power management for laptops? What do you mean by masking it and how do i go about it?

Any help :(?

Have you tried disable Baloo?

balooctl disable

it has been disabled since my fresh install :D, unfortunately that did not help :(.

What happens if you switch to open source nvidia driver? or switching to kernel 5.7.x?

EDIT: I can see in your pastbin that you havn't updatedd to the latest. Any reason?

TLP is not required on a desktop computer. You can mask TP or uninstall tlp.

The forum search feature works quite well for locating uncommon search terms such as:

Mask tlp

I completely removed tlp and rebooted, the problem persists. thanks for the input though, the more possibilities we can eliminate the better!

I chose LTS just to be safe. however, I tried 5.7.9-1 (latest) to no avail.

Bump, I have updated all my attempts...

Bumping is a violation of forum rules:


Posting a single word or useless message (bumping) to attract attention to your thread is not allowed. Do your own research, continue to troubleshoot, post the results, and be patient with the community. If people are reading your thread without answering or offering help, you may try supplying more details, or ask to be pointed in the right direction. Often, the reason for posts remaining unanswered is due in large part to the sparse details in the original post itself, or, the obvious availability of solutions in the wiki, on the forum or on the web, and the community's unwillingness to point out the obvious.

Thank you for letting me know about this rule, I apologize for violating it.

I have done my research, continued to troubleshot, posted updates (or "supplying more details") and also asked whether there is additional information I can supply. I have done everything you listed as recommendations for steps forward.

This has been incredibly frustrating to be honest.

Yes I know it is frustrating as I have experienced it myself.

I found using older kernels and changing my schedulers was a good workaround to eliminating the problem. I would also suggest trying the real time kernels as some real time kernels seemed to not have this problem for me.

Good luck with finding a solution.

Thanks again for your input :)!

1 Like

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.

Forum kindly sponsored by