Noble Dailys: Successes or Failures?

[guiverc]

quiverc -- In that link, I read that as "if there were no partition table" that Calamares leans towards GPT... But to stay like the the other installers, and how I've historically done some of my test-cases to confirm that "as explained a few posts back"... That link does not really differ from what I explained in Post #88, but...--> If you set a partition type and choose "Erase All and Install ___." ...

That the installer should, (in it's rule-based logic)

• if it does not see a partition table, choose a partition table type based on the boot mode (prefer GPT).
• if there is a partition table type set, honor the user's choice of partition table type, and use it.

There are reasons to honor a User's decision, rather than override it.

If it overrides the decision, then, what was told to me about the other installers, that is considered a bug.

---

So if not considered currently as a Bug for Lubuntu, in respect to how Calamares does that... based on that reasoning, I believe it should be considered a Bug. Right? Or do you think there is reasoning for overriding that, and taking the ability to do that away?

I personally, I guess as a technical user, would prefer what you MAFoElffen & Irihapeti have stated, as if I've made a decision (in the past) I would expect an installer to use it; but I have no problem with it due to the "Erase disk" & its creation of a new partition table with that selection. Here it's maybe because I'm still trying to get myself out of habit of using MBR I've I've used since the 80s, but GPT is really the better alternative in modern hardware.

I tend to accept things if I can see logic behind the decision; I do see GPT as better thus had no issue with their preference. When I wanted a MBR partitioned install (in my QA; I try and pretend I'm an imaginary user and create a system that meets an expected need occasionally), I've had no issues (thus far) in creating one; but if I did, then I'd for sure be filing a bug, and myself working out a workaround should a end-user encounter that issue so they're not waiting.

With calamares, a bug report would mean on launchpad (for tracking mostly), then upstream with calamares on github; which is easier given our code on noble isn't old. I've rarely done this, but have found them pretty responsive.

FYI: When it comes to workarounds with partitioning; as end-users are using installers rather infrequently; I've often suggested users use a partitioning tool such as KDE Partition Manager (or gparted) to create their partitioning first; then use the installer to just select where they want things to go (in their already partitioned system). At least for me, as I'm using partitioning tools far more regularly (making changes) than installs, I've always seen this as less risky; this maybe influencing me here.

[MAFoElffen]

I confirmed it: https://iso.qa.ubuntu.com/qatracker/...s/1701/results

I filed it: https://bugs.launchpad.net/ubuntu/+s...s/+bug/2046500

It was the right thing to do. Iripetti and sudodus can go up there and mark that they were affected... So that it's status gets bumped to confirmed.

I figure maybe it might get sent upstream to let Calamares figure out if it is something they want to decide on... Whether as a Bug or as a proposed Enchancement. Or not. Let them decide.

[guiverc]

I confirmed it: https://iso.qa.ubuntu.com/qatracker/...s/1701/results

I filed it: https://bugs.launchpad.net/ubuntu/+s...s/+bug/2046500

It was the right thing to do. Iripetti and sudodus can go up there and mark that they were affected... So that it's status gets bumped to confirmed.

I figure maybe it might get sent upstream to let Calamares figure out if it is something they want to decide on... Whether as a Bug or as a proposed Enchancement. Or not. Let them decide.

Thank you @MAFoElffen, I saw the bug report appear in my inbox; and for reporting on the ISO tracker (meaning it'll appear on random tests into the future, meaning others may test for it, and I get reminded when I see it!).

Thanks also to @Iripetti & @sudodus for their testing & reporting... I've not had much time to QA test recently; and I'm very thankful to see others ahead of me in the cycle testing stats.

[Irihapeti]

Thanks for the report. I've added myself as affected, and subscribed.

[Irihapeti]

Would anyone testing the Xubuntu installer be able to confirm https://bugs.launchpad.net/ubuntu/+s...y/+bug/2045535 ?

The title probably isn't the best. It seems that anything writable connected to the USB bus becomes the default and only choice for installing to when using auto-resize. At the time I made the report, I thought it had to be bootable, but it appears now that it needn't be. I've been able to confirm it in a VM, but because it's still just me, Launchpad doesn't consider that a confirmation. I think this bug has a potential to really mess things up badly for someone.

[TenPlus1]

The one thing that annoys me about the new 24.04 dailys vs 22.04 releases is that the memory use for each desktop has gone right up by 300mb or more, and that's using the same kernel on both just to be sure. What's using all that extra memory ??

[guiverc]

The one thing that annoys me about the new 24.04 dailys vs 22.04 releases is that the memory use for each desktop has gone right up by 300mb or more, and that's using the same kernel on both just to be sure. What's using all that extra memory ??

A recent review of a minimal Lubuntu 24.04 daily install says the following

The minimal install only takes 4 GB of disk space, and its lightning fast. It only uses 419 MB of RAM at idle.

I doubt an equivalent Lubuntu 23.10 install was using only 119 MB.

Personally that 419 MB means nothing to me, almost no-one uses their desktop without additional programs, and starting up some services/processes will mean the idle RAM is higher, but also that starting apps may also be faster. There are always pros and cons with anything, and that includes use of RAM.

If a system I'm using allows me to freely adjust what is started/stopped, and have control over what is running (which Ubuntu does) I'm happy.

fyi: A minimal install option is also something that I don't care about. Myself I find it as easy to uninstall unwanted packages as it is to add those I want; but the world would be boring if we all thought/felt the same.

[sudodus]

Thanks for the report. I've added myself as affected, and subscribed.

+1

[sudodus]

Would anyone testing the Xubuntu installer be able to confirm https://bugs.launchpad.net/ubuntu/+s...y/+bug/2045535 ?

The title probably isn't the best. It seems that anything writable connected to the USB bus becomes the default and only choice for installing to when using auto-resize. At the time I made the report, I thought it had to be bootable, but it appears now that it needn't be. I've been able to confirm it in a VM, but because it's still just me, Launchpad doesn't consider that a confirmation. I think this bug has a potential to really mess things up badly for someone.

I cloned the current Xubuntu Noble iso file to a USB connected SSD and booted into my old Toshiba Satellite Pro C850-19W.

- There is an internal SSD with Lubuntu (since yesterday).
- I connected a 32 GB Sandisk Extreme USB pendrive with an MSDOS partition table and an exfat partition.

When booted in BIOS mode, only the internal drive and the pendrive could be selected for booting.

When booted in UEFI mode, I could boot into the USB connected SSD and run Xubuntu Noble liive. At the partitioning menu I selected 'Install alongside' and only the internal drive was possible to select (the pendrive could not be selected).

In other words, I could not reproduce the behaviour that you describe. Please suggest how I should modify the setup of the test to be able to confirm your bug report.

In your test, is your SD card connected via USB or PCI? If via PCI, I must test in another computer.

---

Edit: In the same Toshiba, in UEFI mode, I replaced the USB pendrive with another one that is bootable with Clonezilla in UEFI mode.

Now there was no option to install alongside, only

- 'Erase Ubuntu Noble' (actually Lubuntu)
- 'Erase disk'
- 'Something Else'

I interrupted the installer before writing anything (I am rather sure about that). Anyway, I repeated booting with the Sandisk Extreme connected and then 'Install alongside' was available again. So I can confirm that Ubiquity is flaky, but I see other symptoms than you.

[Irihapeti]

I need to re-run the tests in UEFI mode to be sure of that, and they are in a QEMU/KVM VM. From memory, it was the same thing. I used a physical USB stick in that instance, because I don't know how to emulate USB things in software.

The original set where I discovered the behaviour was on the Acer travelmate 4740, which is BIOS boot. It has a 500GB SSD in it, and the SD card was connected to the USB bus. I subsequently noted the same behaviour when I inserted a USB stick. The SD card has ext4 file system and the USB sticks have either ext4 or fat32. In all cases, only the external devices were visible for auto-resize installing. (I consider the SD card to be external, even though it sits entirely in the laptop casing when it's attached.) Maybe I should try testing with exfat, seeing as a lot of USB storage comes with that these days.