[SOLVED] Low disk space error at boot (Linux Mint installation)

[Kenneth]

Hi everyone. I am still pretty new to everything Linux related although I've done about a dozen of Linux installations by now.
Further more, as a Servicedesk engineer at a school, I posses decent knowledge of computer issues regarding booting, installations, hardware problems etc... That being said, I haven't found this problem anywhere else on forums and I will try to provide as much information as possible on this issue.

I have issues installing Linux Mint 21.1 Cinnamon on my HP laptop (more hardware information at the bottom of this post).
I'm able to boot Linux Mint via a bootable USB stick and access anything (e.g. browser, tools, menu's, ...).
After a few minutes, an error appears on the top right of the screen, stating "Low disk space - This computer has only 0 bytes disk space remaining". From this point on, I can't do anything anymore (for example opening Firefox or going to the keyboard setting), the computer freezes or crashes entirely (I can still move my mouse around when it freezes). From now on when I boot Mint again, I immediately open the "disk space analyzer" to get more information. I notice that the available disk space goes down quickly, starting from 8GB and after a few minutes being 0 bytes, after which I get the disk space error again and can't do anything. Needless to say, It's also impossible to install Mint in this limited time period. It crashes after a few installwizard screens.

Installing Windows 10 is never a problem, I've done this countless times on this laptop. I have also tried booting with another USB drive and I have also tried installing other distro's than Mint (Ubuntu and Manjaro to name a few), It always comes down to the same issue regarding the disk space.

I don't know if this information is useful: VTx Technology is enabled. I have tried fiddling around with various BIOS settings but that didn't solve anything. I also noticed that there's a TPM device inside (I didn't think that a 2016 laptop would have that).

About the laptop:
HP Notebook 17-x095nb (Energy Star)
Processor: Intel Core i7-6500U CPU @2.50GHz x 2
Memory: 2x 8GB 1Rx8 PC4 - 2400T - SA1 (DDR4)
Graphics card: Intel Corporation Skylake GT2 [HD Graphics 520]
Hard drive: Originally It contained a 1TB SSHD and thought this might have been the issue, but recently I've changed it to a SSD (unsuprisingly this made the laptop much faster), I also tried various SSD's. Right now there's a 256GB SATA2 SSD inside.

I have no clue on how to fix this issue or how I can get more information. It would also be helpful to know if I would need to change some more hardware or if another Linux distro would work.

[gittiest personITW]

Welcome to the forums.
Are you trying to do a format and default install? Single boot (no other operating system)?
Or is there something else we're missing?

[Kenneth]

Welcome to the forums.
Are you trying to do a format and default install? Single boot (no other operating system)?
Or is there something else we're missing?

With the current setup, there is no other OS installed on the SSD. This SSD was formatted before I put it in so all space should be available (although only 8GB seems to get reserved when I boot via USB and I don't know the finer details about this).
My intention is to have a single boot "Mint only" laptop, not a dual boot with Windows or any other OS.

PS: It might be important to note that I usually use Ventoy on my bootable USB stick.

[gittiest personITW]

Just to check - don't take offence....
The 8GB that is reserved is definitely not the USB stick you are using (Ventoy)

Does the SSD show up in BIOS?
Does it show up in Disks in Live Mint?
There is another recent thread that may help that rene suggested to look at BIOS for settings :

The issue is presumably going to be the drive in the BIOS being set to "Intel RST" or some other RAID-type, and/or BitLocker being configured for the drive. If latter you first need to disable that: https://discourse.ubuntu.com/t/ubuntu-i ... d-on/15338. If only former just flip setting (and test if Windows is then still okay...)

[rene]

I would expect that the system's logging itself into oblivion, and maybe as this one: viewtopic.php?f=90&t=388316

If so you'd want to boot the installer with pci=nommconf and would want to configure this on the installed system quickly. You can supposedly check from the Live/Installer system with from a terminal sudo journalctl -f. If indeed, you can add kernel parameters to the Live system boot in the same manner as specified for "nomodeset" in "Solving freezes during the boot sequence" at https://www.linuxmint.com/rel_vanessa_cinnamon.php

[Jo-con-Ël]

Booting linux Min Live USB, would you please provide a Gparted (Menu>Administration>Gparted)/Disks (Menu>Accessories>Disks) screenshot showing complete disk partitioning and/or post back following codes results?

Code: Select all

sudo parted -l

[rene]

Why are you guys all focusing on drives? It's his Live system that runs out of space...

[Kenneth]

I would expect that the system's logging itself into oblivion, and maybe as this one: viewtopic.php?f=90&t=388316

If so you'd want to boot the installer with pci=nommconf and would want to configure this on the installed system quickly. You can supposedly check from the Live/Installer system with from a terminal sudo journalctl -f. If indeed, you can add kernel parameters to the Live system boot in the same manner as specified for "nomodeset" in "Solving freezes during the boot sequence" at https://www.linuxmint.com/rel_vanessa_cinnamon.php

Thanks for the clear answer.

You are definitely right about the logging. I've looked for the logging files in /var/log and noticed syslog and kern.log kept growing in size at a rapid pace. There's also the dead giveaway of my laptop's ventilators working overtime (this only rarely happened when W10 was installed). Running "pci=nommconf" in the terminal and checking this with "sudo journalctl" did not fix the problem; what I got was a never ending log inside the terminal, supposedly this is the logging about which log files are being canceled. This also didn't stop the system writing infinite logs and filling the hard drive.
I've been fiddling around with other distro's and I was able to install Manjaro. The logging and disk space problems still occured but at a much slower pace. Still, this isn't a solution because right now, having installed Manjaro, the logging folders are still growing and it won't be long until my hard drive is full. It DOES give me some more time to research the issue.

I've also taken a quick look at the forum threads you suggested but I still don't have a solution to my specific case. Is there a way to permanently stop the logging on a Linux distro? Or should I try and figure out what the infinite logging is about?

[rene]

Yes, the pci=nommconf isn't a command to be run from a terminal, but a kernel parameter to be specified at boot. In the case of the boot into the Live/Installer USB as per the linked instructions for, in that case, nomodeset instead, but exactly that, just with pci=nommconf rather than nomodeset.

In the case of an already installed system you can specify a kernel parameter as viewtopic.php?t=349669

Of course also first make sure by the mentioned journalctl command or simply looking in those growing log files that it is in fact that "AER: Corrected error received" issue that was linked.

[Jo-con-Ël]

Why are you guys all focusing on drives? It's his Live system that runs out of space...

I'm sorry, but someone is quoting (here) what you said about RAID and sudo parted -l is an easy way to confirm or discard this point w/o getting into computer Setup. Now I see you didn't say that anywhere on this thread or removed it. Good luck.

[rene]

That indeed wasn't from this thread (if you follow the up-arrow on the quote you'll be transferred to where it came from...) and, yes, I also wasn't too sure as to why that was being pulled in here because we're as said looking at the Live system filling up

Any case -- seems that indeed error-logging was the issue; let's wait for poster to be satisfied to have a working solution for that...

[gittiest personITW]

This is interesting if it is the logging issue, and even if it isn't it brings up another question.
I'd assumed (incorrectly it seems) that the amount of 'disk space' on a Live boot was the amount of RAM that was installed.
If this isn't the case, what is the limit to the amount of 'disk space' on a Live boot without persistence?

Next time I do a reboot I'll find out.

[rene]

I don't actually know but would feel it pretty safe to say it's probably half the amount of ram, i.e., the default size for a tmpfs.

[Kenneth]

Hey everyone, it's been a couple of weeks since my last update. A lot of people helped me to troubleshoot the problem and I'm glad to say that I found the solution. I've tried a lot of things such as setting boot flags like "pci=nommconf" and other ways to prevent the laptop from logging at light speed. Those solutions didn't work and the same goes for manually deleting the kern.log and system.log files, it didn't even give me more time.

With the help of one of my teachers I've taken a closer look at the logging and noticed there being a hardware problem. All of these loggings start with "PCIe Bus Error", insinuating that it must have something to do with the network card / network drivers (because a laptop doesn't have a lot of PCI express hardware except for the network card).

The next step was disconnecting the network card and live booting Mint again. Immediately I noticed the fans not making any noise and after that, I was glad to see the available disk space staying the same. Problem solved, except for the fact that I don't have any network (and bluetooth) options anymore. Luckily I searched my house and found another network card that fits into this laptop. Now everything works and I've installed Linux Mint without any problems.

FYI: the dysfunctioning network card is a Realtek 8723BE and I replaced it with an Intel network card. I still don't get WHY the Realtek one wouldn't work...

I've marked this thread as solved.

[rene]

FYI: the dysfunctioning network card is a Realtek 8723BE and I replaced it with an Intel network card. I still don't get WHY the Realtek one wouldn't work...

If you still want to test, if definitively pci=nommconf does't work, pcie_aspm=off stands the next highest chance. However, please double-verify that former doesn't, since it's highest chance.

I.e., you on an already installed system specify the kernel parameter as per viewtopic.php?t=349669 and verify after booting with cat /proc/cmdline that it indeed was specified correctly. You'd try in that order and one at a time,

1. pci=nommconf
2. pcie_aspm=off
3. pci=nomsi
4. pci=noaer

A note is that 4 only disables the ("advanced") error reporting -- but that it's still not necessarily "cheating" since the problem may exactly be said reporting in the first place. 3 also has high change -- but if 2 works that's best.

[gittiest personITW]

the dysfunctioning network card is a Realtek 8723BE and I replaced it with an Intel network card. I still don't get WHY the Realtek one wouldn't work...

Might be this reason
https://easylinuxtipsproject.blogspot.c ... k.html#ID3

[marco_gnudi]

Hi to Everyones,
I would like to share my experience, like the same as Kenneth had.
I have an ASUS F541U Intel Core i5 6198DU up to 2.8GHz, with the network card: Realtek model: RTL8723BE
I tried to install Limux Mint several times but the realtek card always froze the installation with the same error shown by Kenneth.
Once I've replaced the network card by using an INTEL one, Linux installation was fully successful.
For who is interested, I've bought the following card:
https://www.amazon.it/dp/B08SJX6279/ref ... EUBK79VF64
This card is fully compatible with Linux.

[Bjorn Helgaas]

Please, please, please do not consider kernel parameters like "pci=nommconf", "pcie_aspm=off", "pci=nomsi", "pci=noaer", etc as solutions to these problems. They are definitely *workarounds*, but they don't solve the underlying issue, which means that future users will keep running into the same problems.

If you find that one of these parameters is a useful workaround, PLEASE report it to linux-pci@vger.kernel.org. Then we can try to fix the underlying problem so the parameter is not needed for future kernels.

It's helpful if you can include a complete dmesg log from the most recent kernel you have (ideally an upstream kernel) and the output of "sudo lspci -vv". The linux-pci@vger.kernel.org mailing list only accepts plain-text email and limits the message size, so it may be easiest to open a report at https://bugzilla.kernel.org and include the URL in your email.

[gittiest personITW]

Be prepared for an onslaught of emails Bjorn.

I'm sure we all thank you for your keyboard wizardry and hope you feel the love.

[rene]

Be prepared for an onslaught of emails Bjorn.

And more specifically, be prepared for an onslaught of email asking you what a dmesg is, which stream you're talking about being up on, what a mailing list is, and whether or not any of that can be done without needing to open a terminal in the first place...