Printing and identifying disks

Marcos Dione — Thu, 01 Jan 2026 13:41:06 GMT

Another short one. Seems like I haven't completed anything interesting lately, but I have a couple of things in the burner. Also, this is mostly for remembering this myself, because I always spend 30m to come up with this invocation¹. This is another of those you-wouldn't-need-to-read-this-post-if-you-read-manpages-from-time-to-time posts.

I have a laptop that's the web/home/media/backup server and it's the AP. For many reasons, it has 4 disks⁸. I also have like 10 other old disks laying around, waiting for the day I buy a not so old tower to replace the laptop (which is... 13yo). That day, identifying disks will be a necessity. So far, the best tool I have found that gives me an idea what is what is lsblk, but I'm not satisfied with its default output format:

mdione@diablo:~$ lsblk
NAME   MAJ:MIN RM   SIZE RO TYPE MOUNTPOINTS
sdb      8:16   0   1.8T  0 disk 
└─sdb1   8:17   0   1.8T  0 part /mnt/trash/mdione/public_html/Pictures
                                 /home/mdione/public_html/Pictures
                                 /home/backup
                                 /mnt/data
sdc      8:32   0 931.5G  0 disk 
├─sdc1   8:33   0    16G  0 part [SWAP]
└─sdc2   8:34   0 915.5G  0 part /
sdd      8:48   0 698.6G  0 disk 
└─sdd1   8:49   0 686.6G  0 part /home/mdione/public_html
                                 /mnt/trash
sde      8:64   0   1.8T  0 disk 
├─sde1   8:65   0  23.8M  0 part 
├─sde2   8:66   0   1.8T  0 part 
└─sde3   8:67   0  48.8G  0 part

Why? Because I don't care about MAJ:MIN, but also I need more info: partition table type², device model³, and at least available space⁴.

Now look at this⁹:

mdione@diablo:~$ lsblk --all --output NAME,TYPE,VENDOR,MODEL,REV,ROTA,HOTPLUG,RO,SIZE,PTTYPE,PARTTYPENAME,PARTUUID,PARTFLAGS,PARTLABEL,FSTYPE,FSSIZE,UUID,FSAVAIL,LABEL,FSROOTS,MOUNTPOINTS

This space intentionally left blank

It is a lot of info, but:

NAME: yes, and I love the tree.
TYPE: OK, this might be extra, but to this day nvme0n1 still feels like a partition, and not a whole device.
VENDOR: Yeah, OK, this looks useless for internal disks (the other two are connected via USB, see HOTPLUG).
MODEL: yes, please
REV: OK, not that I upgrade disk firmware ever...
ROTA: HDD vs SSD/Flash/NVME
HOTPLUG: mostly, USB, see VENDOR⁶
RO: maybe useful for RO SDs?
SIZE: of course. Unluckily I can't find how to ask for a consistent unit (see the G vs M vs T).
PTTYPE: dos/MBR vs gpt, which is BIOS/legacy vs UEFI.
PARTTYPENAME: might be misleading, see sdb1's and the file system it hosts.
PARTUUID: Debian refuses to allow specifying root partiition by label, and this machine detects disks in the 'wrong' order, rendering it unbootable if I use /dev/sdXY instead, so this is important⁸.
PARTFLAGS: 0x80 is DOS bootable partition, which I need in this non-UEFI machine⁷.
FSTYPE: of course.
FSSIZE: ditto
UUID: see PARTUUID.
FSAVAIL: of course.
LABEL: except for /, I mount by label. Sue me, Debian.
FSROOTS: My disk usage is weird⁸. To me it's imporatnt to know what is mounted where. The many entries with non / FSROOTs are bind mounts.
MOUNTPOINTS: Ditto.

So, in all, this command replaces mount and even mount -t ext4, df -h, fdisk -l, perusing dmesg and maybe more.

Yeah, OK, I made a script out of it, OK? And it's now deployed everywhere (2 machines :) via Ansible. ↩
For when I'm upgrading disks and the new one is not booting or something, like I did recently. ↩
To identify two very similar disks for said upgrades, or make sure that /dev/sda is not the main disk on the machine⁵ but a USB pen drive I want to write a rescue image on; otherwise, I would be /dev[astated]/sad. ↩
Who doesn't want to know this, right? ↩
Since I have a laptop that can handle NVME devices, this has been /dev/nvme0n1, but almost 3y of novelty can't beat more than two decades of precious /dev/sda. ↩
Yes, circular 'see', sue me. ↩
TBF, this machine supports it, but the hassle it means to activate that now is beyond what I'm prepared to fix. ↩
Rationale is like this: I have an internal SSD (not NVME, this machine is too old), this is /; an internal HDD for trash, mainly semi throwaway map tiles mounted in my public_html directory, and it's full; an external HDD, this is for local backup but also serves my pics from a subdir from my public_html; if I have to run and leave everything behind, this is the first/only thing I pick up; and the external NVME on a USB encasing, which is soon to replace the external HDD. ↩↩↩
Chízus, Nikola completely fsck'd it up. Sorry for the text file. And TBH, this does not even fits my terminal on a 170%¹⁰ 3840x2160 27" monitor, but at least looks fine in my navigator. ↩
I'm old, OK? ↩

(More than) Everything you always wanted to know about...

Marcos Dione — Fri, 14 Nov 2025 21:36:59 GMT

From time to time you need to deep dive into a subject and you need a good site to read about it. Over the years I have come across many such sites, and this is my attempt to collect them in a single place. I have revisited the some of the same subjects, but since searching the web has become a worse and worse experience, finding them again has been almost impossible.

I hope to make this list as long as all the sites I want to put here. Also, I'll be using Archive.org's links, to make sure you can find them even if the sites are down.

Certificates and CAs

https://jamielinux.com/docs/openssl-certificate-authority/

Linux memory

https://blogs.oracle.com/linux/understanding-linux-kernel-memory-statistics

UEFI

https://www.happyassassin.net/posts/2014/01/25/uefi-boot-how-does-that-actually-work-then/

Bikes

https://www.sheldonbrown.com/

Sound

[PulseAudio][https://gavv.net/articles/pulseaudio-under-the-hood/]