one big partition :)

[wizard10000]

Well, I finally did it yesterday. Had been kicking around the idea of a single-partition install for quite some time and decided to bite the bullet and do the thing

disclaimer: I'm pretty anal about backups so I didn't risk anything but a little time.

First thing I did was create and activate a 2GB swapfile; except in testing I've never hibernated this laptop and have never seen more than about 50MB of swap used. Made necessary changes to fstab and then booted up a gparted live USB.

Shrunk home partition, removed swap partition and expanded / to take up almost all the free space. Booted the machine in single-user mode, mounted the former /home partition to /mnt and moved my home directory back under /home. chown -R my home directory just to make sure things would work correctly.

Booted gparted again and removed the partition formerly known as /home and resized / to use the entire disk.

Okay, first *real* boot. Kernel complains about not being able to find resume location (didn't see this one coming) so deleted /etc/initramfs-tools/conf.d/resume, ran update-initramfs -u, rebooted the machine and the complaints stopped

Everything works fine. p.H. and I had some discussion earlier about swapfiles being a hack (and they are) but I think this particular hack is here to stay. So far, so good

[p.H]

I am not going to comment on the swap file, you know my opinion.

Just a note : if the / partition was located before the /home and swap partitions, you did not need to boot Gparted live and could have performed all the operations from within the installed system.

[wizard10000]

...Just a note : if the / partition was located before the /home and swap partitions, you did not need to boot Gparted live and could have performed all the operations from within the installed system.

I know how to resize an unmounted partition and filesystem; would be interested to learn how to resize a mounted root partition. If you have the time a quick overview would be greatly appreciated

[p.H]

Resizing a mounted partition and filesystem (online resize) is done in the same way as resizing an unmounted partition and filesystem.
What you cannot do is reduce a mounted ext* filesystem (but you can extend it) or move (i.e. change the starting position of) a used partition.

[wizard10000]

Resizing a mounted partition and filesystem (online resize) is done in the same way as resizing an unmounted partition and filesystem.

I did not know this was a thing. Gonna go read about it now

thanks -

[wizard10000]

I am not going to comment on the swap file, you know my opinion.

And I'm gonna beat you to saying "I told you so"

Reasonably light swap usage (I was using easymp3gain against probably 80GB of music) when all of a sudden *everything* stopped working. On reboot the only way I could log into a console was in single-user mode. Looks like among other things, bash was a casualty,

I haven't broken a Debian install badly enough to justify a reinstall in probably eight years, so this was a humbling experience. The fact that the current mini-iso is broken didn't help matters at all, so I ended up using a netinstall image, which worked fine.

I'm about 80% back up and running, got a few more applications to install. BTW (and you don't get to holler at me for this either), getting gksu working on a fresh Sid install was a bit of a pain in the backside

[CwF]

getting gksu working

Why? That's like installing a deadbolt lock in the bathroom door.
After I took out gksu from stretch the only thing I missed was the cool root terminal icon... I use it for a root terminal button not provided by gksu...You can enable the user with sudo to nix the password, thereby eliminating the need for password entry, thereby not needing gksu. For most common needs pkexec is working now.

Did you figure the crash had something to do with the swap file?
I have mentioned I use either a partition or nothing. I've thought of adding a swapfile to the nothing option, but so far survive with nothing. Is the suggestion a swapfile is less stable than no swap at all?

[wizard10000]

Why? That's like installing a deadbolt lock in the bathroom door.
After I took out gksu from stretch the only thing I missed was the cool root terminal icon... I use it for a root terminal button not provided by gksu...You can enable the user with sudo to nix the password, thereby eliminating the need for password entry, thereby not needing gksu. For most common needs pkexec is working now.

Because pkexec is a bigger pain than gksu is

On a server I wouldn't need it, actually I really don't need it at all but the point, click, give it a password thing works pretty well for me in a graphical environment

Did you figure the crash had something to do with the swap file?
I have mentioned I use either a partition or nothing. I've thought of adding a swapfile to the nothing option, but so far survive with nothing. Is the suggestion a swapfile is less stable than no swap at all?

p.H. and I had a short but interesting discussion about swapfiles vs. swap partitions in this thread. I decided (against his advice) to try a swapfile and I can't say for certain that the swapfile contributed to the explosion but since I have a shiny new Sid install I also have a shiny new swap partition

I've done this task (normalizing about 80GB of .mp3) on this machine probably a dozen times; the process generally takes about 50MB of swap. This is the first time the machine has written anything to the week-old swapfile and only about 35MB was in use when the machine went completely unresponsive (cursor would move, nothing else worked; switching to another console just gave a flashing cursor).

On reboot everything looked normal until you tried to log into a console, which failed no matter which user - login failed to execute /bin/bash: exec format error.

Anyway, about an hour of googling didn't get me much except answers for a chroot that I wasn't doing, so since I'm pretty anal about backups I just decided to blow the thing away rather than fix it. Took me four hours but the machine is running about 50MB leaner at idle than the years-old installation I had before. I'm about 98% finished setting it up, still have to do some kernel autoremove magic (I keep two).

Anyway, sorry for being bloggish - I can't say for sure it was the swapfile but that's really about the only thing that had changed on the machine.

cheers -

[CwF]

; switching to another console just gave a flashing cursor).

...that does mean you were still alive?!

Anyway, ya I was in that thread. I guess I'll skip the swapfile question then. I've been intending to try busters' zram-tools. I will try it in a memory restricted bare metal test with a swap partition first, then a zero swap vm, and try to mention it somewhere. My impression is it still would like to have some disk space in a swap partition. I'm not sure it's intended to forego swap altogether.

[Head_on_a_Stick]

I ran a partionless multi-boot system on an SSD for a while using btrfs subvolumes & GRUB. Prone to breakage but I liked the minimalism.

[p.H]

p.H. and I had a short but interesting discussion about swapfiles vs. swap partitions in this thread. I decided (against his advice) to try a swapfile and I can't say for certain that the swapfile contributed to the explosion

To be honest I would not expect a swapfile to cause such breakage. Swapfiles have been used on ext* for ages, so I'd expect it to be quite stable. It seems that some distributions such as Ubuntu and Mint have switched to using a swapfile instead of a swap partition (to my despair).

[wizard10000]

To be honest I would not expect a swapfile to cause such breakage. Swapfiles have been used on ext* for ages, so I'd expect it to be quite stable. It seems that some distributions such as Ubuntu and Mint have switched to using a swapfile instead of a swap partition (to my despair).

Yeah, I'm not ready to blame the swapfile for the meltdown. Could have been just a perfect storm of breakage but I did get a shiny new Sid install that uses about 50MB less RAM than my years-old installation did

I had some issues with systemd NFS automounts hanging on shutdown and that seems to have been resolved as well - my override file to stop *.automount worked, but I need for stuff to work right instead of adding workarounds. Anyway, I keep pretty great backups so all I really lost was about four hours of my time and got a machine that runs considerably better than it did before I started messing with it

cheers -