Dusting off some Kindle books, aka this should be easier than it is ►


2021-02-05 📌 Project Zebra: Proud of machines that think, suspicious of people who try to

Tags All Linux Tech Personal

This entry is part of my Project Zebra series covering migration to Linux for personal computing use.

Title reference: an abridged W40K aphorism.

Not that my desktop PC has seen much use except with a small side monitor for email and videos, but I was running out of headroom with the drive in it and it was time to stick a new one in so that I can pick up with projects at some point. The existing 1TB drive is 70GiB system, 10GiB swap, rest data (NTFS partition) and I wanted to split the new 2TB drive into 100GiB system, 20GiB swap in readiness for sticking 16GiB RAM into the box (so that it can all be under hibernated under worst-case circumstances) and the remainder data, still NTFS... which I know is a questionable decision, but what I'm used to in term of case-sensitive behaviour and what numerous backup devices are expecting filenames to conform to.

What follows is a slightly idealised running order and not me forgetting and having to google things. USB adapters and buses being what they are, it's best if the device being written to is connected directly by SATA cable, so after putting the new drive into the PC and working from a bootable Kubuntu 20.04 LTS flash drive, this would have gone:

- sync contents of data partition to new location
- copy system partition using KDE Partition Manager
- update /etc/fstab UUIDs and update UUID for resume in /etc/initramfs-tools/conf.d/resume
- follow https://help.ubuntu.com/community/Boot-Repair to install boot-repair to sort block sizes and reinstall grub, then edit /etc/default/grub again to turn off the splash screen, unset GRUB_TIMEOUT_STYLE and set timeouts to something sensible, set the resume UUID and update its config

So the top of my config file is;

GRUB_DISTRIBUTOR=`lsb_release -i -s 2> /dev/null || echo Debian`

Noticing high CPU usage when I came to move a load of virtual machines off an external drive to the larger main one, I looked into it and have added big_writes to the fstab options for the data partition https://unix.stackexchange.com/questions/536971/disadvantages-of-ntfs-3g-big-writes-mount-option

After that I created a new encrypted volume on a new external 2TB device to shuffle backups around onto, finding it simpler to do that with the original Truecrypt and notes for command line options I'd left myself previously than with Veracrypt (I'm sticking to TC volumes for consistency). That gets me back to full backups for the NAS but my off-site backups have been in the office for about a year.

Getting around to maxing out the RAM was also a bit of a no-brainer as I do tend to run with a lot of browser tabs, use VMs, etc. It's not a need but memory usage can escalate quickly in some scenarios. Sadly the cheap shitty Lenovo Yoga 300 laptop I've got has onboard RAM that can't be upgraded with the 2x4GiB left over. Maybe I should offer it to work, there are still machines of similar vintage that only have 4GiB knocking around.

The M92p model of Lenovo ThinkCentre has been around for a while, but they're solidly built and the i5-3470T runs very comfortably even considering graphics acceleration is off in applicable apps due to intermittent instability under both Windows and Linux. The processor spends most of its time stepped down and the box attached to the back of a monitor with lots of cables into it.

Some snippets… I've bound htop to Win+H because I find KsysGuard more use as a task manager than a processor usage monitor.

For reference, Nemo saves its per-folder view settings in ~/.local/share/gvfs-metadata (my home user folder had gotten its settings stuck despite a setting being on that's supposed to override them).

NG wrote nice bits on KDE stewardship and not screwing users over, and culture;

Recording specific quality with youtube-dl:
youtube-dl -f 'bestvideo[height<=720]+bestaudio/best[height<=720]'

Recorded some youtube videos with unicode emoticons in the filenames? You might find 'detox' helpful – a small command line utility in the Ubuntu repos;