Strange MX Update Issue

I don’t usually post for help mid week, but I am stymied by events since this past weekend with my MX Linux install.

First, last Friday, I made a snapshot of it (took over an hour) and not only saved that to my ext HDD, but put it on its own SSD as it is my daily driver that I depend on.

So over the weekend, I decided to break out an old game manager I had not used for some time and had to modify /etc/security/limits.conf with the following:

username hard nofile 524288

Then I had to reboot. When MX loaded the desktop, none of my backgrounds would load, so I looked in file manager and found that none of my ext media was mounted. It has always automounted before.

Then I saw 24 updates. So updated. One was Waterfox, and I had not closed it prior to updating.

After updating, I went back to the game manager and opened it and it was as if I had never done all the work of installing the app needed for one game. It repeated the info that I needed to modify the same security file.

I opened the file and none of my changes remained. So I did it all over again. Then had to reboot.
Again, no desktop images (due to no automount of where they are stored) and my protonmail bridge was “signed out” and that has never happend before. Every reboot, I am auto logged back in so that my email app will work with proton mail.

Then I went back to the game manager, and same thing. Start all over.

But most importantly, 24 updates showed up again. Now memory is not so good, but I swear they were exactly the same as I already did. So this time, took a screenshot. Even used autoremove in terminal to get rid of unneeded packages.

Then I opened up Waterfox (remember that Waterfox was in the updates from v. 6.9 to 6.10) and “about” showed I was still on 6.9. PLUS it opened previous tabs not last open when I exited.

This was getting weird.

Today I opened my Obsidian app to access my business notebook, and it was as if I had just installed it. No notebooks available, it wanted me to create them from sync. No, no, no. My notebooks reside on my internal drive. So I looked. The folders for them were gone. Thankfully, I had just made a “backup” at the end of May because “sync” is not a backup. I usually do this monthly and store it on ext HDD. PLUS I have them on every computer where Obsidian stores your vault locally on that computer.

The backup notebooks remain on that ext HDD. But I do not know what happened to the local copies on my SSD.

Later, 24 updates appeared. I opened the package manager and my screenshot, and they were exactly the same updates, even the same reference to those needing “autoremove.”

So what exactly can be causing this? Surely modifying the security config file would not do all of this? Besides, it does not stick–the change I make after saving are gone after each reboot.

I do not even know where to begin. But I do know that I cannot continue to reboot and have nothing as it was, including signed in accounts, folders missing, updates that did not stick, etc. I even lose all of my history in terminal as there is nothing after reboot.

And just before posting this, I updated for the 4th time, those same updates and yes…they are back.

Sigh…
Sheila

2 Likes

I have done MX updates this week. No such trouble, but of course I have a different set of apps to update. The security updates would havdebeen the same for both of us.

An update does not normally interfere with config files or tabs.

That game manager that you modified the config file for… is it installed via the package system?

Are you using session saving? That could make the desktop go back to a previous saved state.

https://serverfault.com/questions/300749/apt-get-update-upgrade-list-without-changing-anything

Can you do a ‘pretend’ update of those 24 updates and set some extra parameters so you can see what they would do?

1 Like

I have always used it ever since a first issue arose last year. I currently have two saves and have been using the latest one for the past couple of months.

I do not know how to do that with “parameters” but this last time, instead of using the pkg manager in MX where I usually do the updates because it shows you terminal output, I saw those 24 sitting there and went straight to CLI and did both flatpak and system updates there. No output errors. But then, there have never been any so long as I don’t reboot.

I searched web and have found nothing relevant to this recently in MX. Some older issue from 2019, but my next step, when I have time, is to use that snapshot I made with the MX tool last Friday before all of these issues began.

Yes, installed from MX pkg manager.

For now, forget about the game, who has the time now? I work all week and was only looking to relax while off for my 3-day weekend. :smiley:

But I need to find out why updates are not sticking. There was nothing really about them that stood out to me. I was going to upload the list of what keeps reappearing, even though I wonder if it matters. Any updates, no matter what they are, would not stick because something else is wrong? BUT, since I rebooted, all of those images are gone. I store all of my personal stuff on the ext /home folders, but things like screenshots I allow to be saved on the internal home folders. Guess what? There is nothing in any of them now.

Sheila

1 Like

So I dug through and found the log for the updates I just did via CLI:

Install: libqt6webenginecore6-bin:amd64 (6.4.2-final+dfsg-1, automatic), libqt6quickwidgets6:amd64 (6.4.2+dfsg-1, automatic),
libqt6quick6:amd64 (6.4.2+dfsg-1, automatic),
sse3-support:amd64 (15.1, automatic),
libqt6qml6:amd64 (6.4.2+dfsg-1, automatic),
isa-support:amd64 (15.1, automatic),
libqt6webengine6-data:amd64 (6.4.2-final+dfsg-1, automatic),
libqt6webchannel6:amd64 (6.4.2-1, automatic),
libqt6printsupport6:amd64 (6.4.2+dfsg-10, automatic),
libqt6webenginewidgets6:amd64 (6.4.2-final+dfsg-1, automatic),
libqt6webenginecore6:amd64 (6.4.2-final+dfsg-1, automatic),
libqt6qmlmodels6:amd64 (6.4.2+dfsg-1, automatic),
libqt6opengl6:amd64 (6.4.2+dfsg-10, automatic),
libqt6positioning6:amd64 (6.4.2-1, automatic)
Upgrade:
libdjvulibre-text:amd64 (3.5.28-2, 3.5.28-2.1~deb12u1),
lo-main-helper:amd64 (25.07.01mx23, 25.07.03mx23),
thunderbird:amd64 (1:128.11.0esr-1~deb12u1, 1:128.12.0esr-1~deb12u1),
libjxl0.7:amd64 (0.7.0-10, 0.7.0-10+deb12u1),
waterfox:amd64 (6.5.9-0+4.1, 6.5.10-0+1.1),
libdjvulibre21:amd64 (3.5.28-2+b1, 3.5.28-2.1~deb12u1),
djvulibre-bin:amd64 (3.5.28-2+b1, 3.5.28-2.1~deb12u1),
mx-packageinstaller-pkglist:amd64 (25.07.01mx23, 25.07.03mx23),
mx-viewer:amd64 (24.3, 25.7),
linux-libc-dev:amd64 (6.12.33-1~mx23ahs, 6.12.35-1~mx23ahs)

Now, I am going to reboot and see if those reappear since this is the first time I will be rebooting after updating myself in terminal.

Sheila

1 Like

Okay that was a worthwhile effort. Because this time, I went into BIOS to check and there was a new unknown entry for booting: “Linpus Lite” What in the world is that?

And below that was my MX. So I moved it up and booted and now all of my folder contents are back and my protonmail bridge was logged in. And when I chose the session saved, it said July 6 instead of today, like when I chose earlier after reboot.

So somehow some “lite” booted and my GRUB menu did look a bit weird font-wise in that boot, but I definitely noticed this boot, my GRUB was back to smaller fonts.

All of my terminal history is back in this session as well. AND my Obsidian notebooks returned to their normal place and opening the app was back to what it had been before.

Now I have to see what that is and why it is there and most of all, how to get rid of it.

Sheila

3 Likes

There has to be something wrong with an update to X11 (or maybe Xfce).
Your apps like Waterfox are firing up, but not looking at the config files… so they look like a first start.
You can get that same effect if you put an app in firejail… it does not see the config files.
I just noticed Waterfox is a Flatpak… ie a container.
Maybe Flatpak has some role in this issue?

On Linpus Lite:
https://askubuntu.com/questions/1363886/why-does-ubuntu-show-as-linpus-lite-when-i-am-in-the-boot-manager-is-this-som

It seems like it is some sort of joke name that may be programmed into your BIOS.
if you ask me, it is in very poor taste confusing people with such rubbish.

1 Like

I did not find your forum topic but another one where it was said that Samsung USBs can sometimes be named that in BIOS. Well I think I figured it out from there.

Remember I said I made a snapshot in MX of my current system and put it on the Samsung SSD that is USB-C attached. Well that is what is booting by default because I plugged it in thinking I would boot into it later to see that it worked. And that is what has been named “Linpus Lite.”

Well it works alright, but it is my system without those data files, I guess, like my Obsidian notebooks.

It’s great that I now know I can boot into my MX from there, but obviously I need to add some personal files or tweak some settings.

The only thing that I don’t understand, if that is an SSD with the OS on it, why can it not be updated? Because it is a snapshot? So the snapshot is only to be used to put back on my internal drive and from there boot from that?

Sheila

2 Likes

A snapshot would be like an rsync copy of the OS… before you can boot it you need to fix /etc/fstab and /boot/grub/grub.cfg.
The boot process will boot whatever it finds as the root filesystem in fstab. … if that is the originzl filesystem rather than the copy, it will boot the original rather than the copy. Very confusing results follow.

1 Like

Well, technically, my laptop has been booting it regardless of fstab being edited, but as you say, it’s a mess when booted because it is a copy.

I gather because I put it on that SSD (which is a specific Samsung ext SSD you can treat just like the internal one), BIOS decided to boot it.

Of course this makes sense now. Don’t know why I thought I could just put an rsync on there and boot just like the internal installed MX does. Rsync copy is just like a timeshift copy and it will boot, but I was trying to get the snapshot on it’s own separate drive (not the ext HDD which is slower) so that I would have a “ready to go” solution if my daily driver failed.

I had previously done just that last year when we were trying to get the Garuda on MX GRUB and I kept it there for a while but as time moved on and we had succeeded in keeping MX controlling GRUB, I wiped it and thought I should again have one on there. But I remember we used Clonezilla to put that on the ext Samsung SSD–not rsync.

Thanks!
Sheila

1 Like

I disagree… it has been booting what fstab points to

Does that mean you can see what has happened?

It is not quite 'ready to go '. I may have misled you there when I called it that. You do need to fix fstab. It is a simple fix… make sure / is set to where the copy is… not to where the copy came from.

1 Like

I think the BIOS itself picked up that drive, named it Linpus Lite and put it in 1st boot order. Because I checked my fstab and using blkid, compared what is listed:

myviolinsings@MX-Acer:~
$ sudo blkid
/dev/nvme0n1p5: UUID=“406d8c92-eb3f-4642-8532-fe14c97261cd” UUID_SUB=“1fb52a33-a158-47ae-aaf1-bf0dab2e7cbf” BLOCK_SIZE=“4096” TYPE=“btrfs” PARTUUID=“a7ac825b-dfc3-4d58-bae9-a7fbb8ee2a07”
/dev/nvme0n1p1: LABEL_FATBOOT=“EFI-SYSTEM” LABEL=“EFI-SYSTEM” UUID=“8C3B-BE28” BLOCK_SIZE=“512” TYPE=“vfat” PARTLABEL=“primary” PARTUUID=“f637a525-4b4d-4d41-8629-442ab14a1385”
> /dev/nvme0n1p2: LABEL=“rootMX23” UUID=“0e7fef02-d7f1-48d6-b9e7-56d51ba1b133” BLOCK_SIZE=“4096” TYPE=“ext4” PARTLABEL=“MX LINUX” PARTUUID=“97920760-0e8d-46ff-b320-e43a17972466”
/dev/sdb2: LABEL=“Linux Acer ES” UUID=“1084e7b1-a909-4e58-9600-d3087c81956d” BLOCK_SIZE=“4096” TYPE=“ext4” PARTLABEL=“Linux Partition” PARTUUID=“b9d229ba-46af-4221-a340-64f8dc8951ad”
/dev/sdb1: LABEL=“easystore” BLOCK_SIZE=“512” UUID=“4C4E29694E294CD0” TYPE=“ntfs” PARTLABEL=“Acer EasyStore” PARTUUID=“9ff48cd8-684b-44a2-b378-170afbb650c5”
/dev/sda2: LABEL_FATBOOT=“LIVE-UEFI” LABEL=“LIVE-UEFI” UUID=“FA44-C8E9” BLOCK_SIZE=“512” TYPE=“vfat” PARTLABEL=“primary” PARTUUID=“abd2cc3d-2430-4187-8603-b87096973e24”
/dev/sda1: LABEL=“Live-usb” UUID=“af1a06bb-85e4-468c-84ef-767847cab76d” BLOCK_SIZE=“4096” TYPE=“ext4” PARTLABEL=“primary” PARTUUID=“a97f6c09-4f53-471f-b094-6f849038ac8a”

The nvmen1p2 is my internal M.2 drive that boots to MX GRUB and from there I can choose Garuda (on nvme1p5) or MX.

My fstab is showing that UUID as the boot device:

UUID=0e7fef02-d7f1-48d6-b9e7-56d51ba1b133 / ext4 noatime 1 1
UUID=8C3B-BE28 /boot/efi vfat noatime,dmask=0002,fmask=0113 0 0
/swap/swap swap swap defaults 0 0

What I had before on this ext SSD was created from using Clonezilla to lay the image of my current install onto that bootable SSD. At that time, you could tell no difference between the two.

I used the MX Tools “snapshot” feature this time and had put that on the drive.

At least I know that in BIOS, that Linpus Lite is still there and since I changed the boot order myself and put MX back on the top, it will work fine. And should my BIOS ever try that again, I will know what is up. :smiley:

Sheila

3 Likes

Yes, the BIOS picks the boot drive (ie where to get grub) , the grub menu picks the partition to get the initrd image, but fstab (and maybe grub.cfg) picks the root filesystem partition ( ie where to get linux)

2 Likes

I searched for efibootmgr from mxlinux.org and found this: UEFI Manager – MX Linux

Check if the linbus lite entry shows there and delete it. I’m more used to work with cli than GUI but the instructions are great on that link. Here’s Gentoo wiki on the same issue efibootmgr - Gentoo wiki

Linbus lite was some kind of closed source os based on Linux IIRC

2 Likes

Thanks for that @ihasama. I did not know MX had that. Here is what mine showed:

I have never been sure why there is an MX Linux and an mx in my BIOS, but also, that ext SSD is showing up as Boot 0002 while Linpux lite is 0000. Curious.

I am removing Linpus and then will see what UEFI Boot manager displays upon reboot.

Sheila

2 Likes

I think one is for a grub boot, and the other is for an UEFI boot.

1 Like