Hello:

Just a (belated) comment / heads-up:

My Devuan Daedalus system runs on a box with a 120GB SSD and 8GB RAM.
As my conky swap file use always read nought, I decided to reduce it to 2.6GB just in case some application looks for a swap file.
I plan to reduce it further should I need more space in / or /home.

Long after that, I set up a tmpfs like you suggested.
It ran without a hitch till a couple of days ago when I had a system lock-up while streaming a video.
Something I had never encountered before and made me fear for my system's SSD.

After making a full Clonezilla image of my system drive, I had a look (among other things) at df -h and found that it was set up with a capacity of 3.9GB, something I was not aware of:

$ df -h
--- snip ---
tmpfs           3.9G   18M  2.0G   1% /home/groucho/.cache/mozilla/firefox
--- snip ---
$ 

This is 50% of the available physical system RAM in my box.
And the default value tmpfs claims when it is set up without specifying the desired size.
In this specific case, it was more than the amount assigned* to /dev/shm:

$ df -h | grep tmpfs
tmpfs           788M  800K  787M   1% /run
tmpfs           5.0M  8.0K  5.0M   1% /run/lock
tmpfs           2.1G     0  2.1G   0% /dev/shm
tmpfs           3.9G   20M  2.0G   1% /home/groucho/.cache/mozilla/firefox
tmpfs           788M  8.0K  788M   1% /run/user/1000
$

* apparently limited by the amount of available RAM in a system

Suspecting some sort of conflict (?) somewhere, I did some searching and found this tidbit which may be of use.

I have now set it up with a specified size of 2GB:

$ cat /etc/fstab
--- snip ---
tmpfs     /home/groucho/.cache/mozilla/firefox  tmpfs nosuid,nodev,noatime,user,uid=1000,size=2G  0  0

$ df -h
--- snip ---
tmpfs           2.0G   18M  2.0G   1% /home/groucho/.cache/mozilla/firefox
--- snip ---
$ 

I really don't really know if the two lockups I had were caused by my system using a 3.9GB tmpfs but I'll report back if I get another lock-up.

Best,

A.

Hello:

Let's see ...

Old               ☑   
Geezer         ☑  -> actually, one of the US english slang terms for 'old'.
Pure ALSA   ☑

Yes, add me to the list.

Nope.
Nocando Miss ...

Just using pure ALSA is not enough to qualify for that list.

Best,

A.

Hello:

I could have sworn I did a copy/paste of the original text by fsmithred.
Evidently not.

In other words, a syntax error (?)

Thank you for taking the time to explain all this.
Learn something every day.

And I should have checked ...  8^°

But it worked, I am no longer getting that line in my dmesg:
ie:

--- snip ---
[   24.068315] PM: Image not found (code -22)
--- snip ---

Thank you very much for your input.

Best,

A.

Hello:

I had deleted the edited file and returned to the back-up I saved.
So I re-edited it now.

$ cat /etc/initramfs-tools/conf.d/resume
# changed 06122024
# see https://dev1galaxy.org/viewtopic.php?pid=53266#p53266
# this box does not use hibernation / resume

RESUME=none

# RESUME=UUID=f0187ff0-be52-4bbc-9461-40f744554b85
$ 

But now it works.

# update-initramfs -u
update-initramfs: Generating /boot/initrd.img-6.1.0-28-amd64
live-boot: core filesystems dm-verity devices utils udev blockdev dns.
#

I'm sure I had written / spelled it correctly.

But if it did not work, there was obviously something amiss.
Maybe the added comments?
Not so sure now.  8^/

Thank you very much for your input.

Best,

A.

Hello:

This post caught my eye.

Don't want to hijack the thread but seems relevant.
So just one question, if I may.
Please let me know if it warrants a separate thread.

That is exactly what my box has and the UUID is the swap file.
But it happens that I do not keep my system on all day, just shut down and reboot.
ie: there is no hibernation image present, never set it up to hibernate.

At some point my dmesg printout reads:

--- snip ---
[   24.068315] PM: Image not found (code -22)
--- snip ---

Q:
Does the system look for a hibernation image at boot time because the line in /etc/initramfs-tools/conf.d/resume is not set to =none and there is a valid (swap file) UUID present?

If I edit /etc/initramfs-tools/conf.d/resume and set RESUME=none, on updating I get this:

# update-initramfs -u
update-initramfs: Generating /boot/initrd.img-6.1.0-28-amd64
/usr/sbin/mkinitramfs: 5: /etc/initramfs-tools/conf.d/resume: none: not found
I: The initramfs will attempt to resume from /dev/sdb3
I: (UUID=f0187ff0-be52-4bbc-9461-40f744554b85)
I: Set the RESUME variable to override this.
live-boot: core filesystems dm-verity devices utils udev blockdev dns.
# 

The system looks for a swap file and not finding one with a UUID of none, finds one and uses it.
What to do?

Best,

A.

Hello:

@Ron
See https://www.linux.org/threads/clearing- … ost-250695

That said, I run a Chimaera VM with pi-hole and unbound inside my Daedalus box.
There is a DNS cache but it is mine.
It can be flushed though the UI.

Best,

A.

Hello:

I used to think more or less along the same lines.
But after reflecting on the matter (on and off fashion) I came to the conclusion that he may be an asshole, an idiot of sorts too.

---> But he is quite clearly a capable developer / programmer. <---
Just why do you think he is working for YouKnowWho?

I'd try to avoid continuing to underestimate Poettering, at your+everyone else's peril.

What I once labelled as being an asshole+idiot was the result of my frustration at what was going on.
But then I realised that it was my reaction (questionable and angry ad hominem manner) against the fact that he had evidently been working to infect Debian with the infamous systemd* for quite a long time, maybe years.

You do not come up with such complex shit overnight, no matter how capable you are.

After that, he outed himself by openly going to work for MS, where he now continues with the classic +30 year old Embrace Extend and Extinguish process we are all familiar with, process that he surely started from inside Debian much before systemd was ever heard of.   

* nothing but an MS registry for the Linux ecosytem

And here we are.

Best,

A.

Hello:

The changelog file in pluma_1.26.1-2.1_amd64.deb does not seem to mention it (?): 

### pluma 1.26.1

 * Translations update
 * Fix out-of-bounds write
 * Fix double activation of extensions
 * pluma-plugins-engine: fix memory leak

### pluma 1.26.0

 * Translations update
 * filebrowser-plugin: fix segfault on open-terminal action
 * update copyright to 2021
 * build: fix shebangs
 * warning: declaration of  shadows a global declaration
 * pluma: Use EXIT_SUCCESS macro instead of int value (portability)
 * help: update help for sort plugin
 * sort plugin: Port sort plugin to the new GtkSourceView api.
--- snip ---

Cannot be tested in Daedalus; GDebi won't install -> non-satisfiable dependency: gir1.2-pluma-1.0 (=1.26.1-2.1).

Best,

A.

Hello:

Sure, no problem with that.

Thanks for that info.

Of course, always.

But ...
On one hand, I have not heard from anyone (yet) with the same issue and on the other, I have not been able to reproduce it.
Not has it happened again.

What I know for sure is that I had not ever seen it happen before the last two Firefox *esr updates.
ie: the versions without the upper left hand View recent browsing across windows and devices icon.

Can that new crap feature be disabled?

And then it may be a fluke, a hardware problem or worse, allucinations. 8^°
No idea, but I don't like it.

Thanks for your input.

Best,

A.