I'll note, lsblk can return a heck of a lot more data than it does by default (and nvme drives show up there). lsblk -H will list for your system, and you can specify columns. You can also adjust output.
I guess with this in mind, I'm curious how this is different?
tanelpoder · 2h ago
Hi, yep lsblk targets a wider area of functionality, like showing mountpoints, device UUIDs, while lsds focuses only on block device settings.
Maybe the latest Linux versions have lsblk versions that support these columns, but in RHEL9 at least I don't see equivalents to lsds'es WBT_LAT, QDEPTH (not the same as lsblk's RQ-SIZE), WCACHE, FUA and some others. But these 4 are which I regularly need (especially when troubleshooting a yet another slow fsync() issue etc). I did and do use lsblk all the time too, but still end up catting and grepping various additional files and correlating the results, sometimes on systems with 100+ multipath block devices.
The other reason was that I wanted a tool that shows me where it gets these values too (for myself and sometimes for explaining stuff to others).
Edit: That being said, it shouldn't be hard at all to add the said extra fields to lsblk too.
bitbang · 4h ago
Very nice, needs option for json/jsonl output.
tanelpoder · 4h ago
Thanks! Yep I was thinking of doing that next, will be very easy as under the hood the data is stored in Python dictionaries.
trillic · 1h ago
Rewrote most of the functionality in C as an exercise
I always wanted the /dev/zero character device driver, which you can map into memory to clear it, or use as an infinite source of nulls, to use the minor node number as the value that got mapped into memory or produced, so you could make an infinite source of beeps with:
mknod /dev/seven c 1 7
I wonder what would happen if you made a /dev/seven device in your http servers public_html directory? Would it dutifully serve it up?
Better yet, support for utf-8 unicode, so you can make an infinite source of poo emojis.
The "Everything Is A File" philosophy should be taken to its logical conclusion.
Maintaining an AUR package can be great fun and an instructive glimpse into what FLOSS maintainers go through.
tanelpoder · 3h ago
I just added a little comment/errata regarding the NVME_QDEPTH column to the post (search for errata). I should probably rename that column to emphasize that (for now) it’s the Linux nvme module level max QD and not the hardware one (it’s complicated…)
I guess with this in mind, I'm curious how this is different?
Maybe the latest Linux versions have lsblk versions that support these columns, but in RHEL9 at least I don't see equivalents to lsds'es WBT_LAT, QDEPTH (not the same as lsblk's RQ-SIZE), WCACHE, FUA and some others. But these 4 are which I regularly need (especially when troubleshooting a yet another slow fsync() issue etc). I did and do use lsblk all the time too, but still end up catting and grepping various additional files and correlating the results, sometimes on systems with 100+ multipath block devices.
The other reason was that I wanted a tool that shows me where it gets these values too (for myself and sometimes for explaining stuff to others).
Edit: That being said, it shouldn't be hard at all to add the said extra fields to lsblk too.
https://gist.github.com/grahameger/2507019334f07036f84080a87...
mknod /dev/seven c 1 7
I wonder what would happen if you made a /dev/seven device in your http servers public_html directory? Would it dutifully serve it up?
Better yet, support for utf-8 unicode, so you can make an infinite source of poo emojis.
The "Everything Is A File" philosophy should be taken to its logical conclusion.
Maintaining an AUR package can be great fun and an instructive glimpse into what FLOSS maintainers go through.