I wrote the maintainer and he told me he no longer has the time to take care of it

(there is a small bug since the package binaries were moved from /usr/bin to /usr/local/bin)

but in general it can be other packages in the future

Is there another active irc chan for freebsd-arm or mostly discussed here?

Trying to find other users/devs working on pwm/power/te,p drivers for the bcm soc

s/te,p/temp

Tenkawa: efnet#bsdmips maybe ?

Ahh good idea

wiki.freebsd.org/Torrents is lagging... NO 14.2 yet -- get your shit together :P

I thought I was forgetting one.. thanks

HER: I'm actually quite pleased how well it runs so far on the RPI5 with NVMe so far.. just needs some tweaks

HER: did you the see the package?

uskerine: no

I wrote it above, since you asked I thought you would want to see which one was it

uskerine: there is porters handbook for that, miles long and very good

metric miles

:p

I am opening it, it is very difficult to port a package provided that it compiles out of the box?

but what would be rsh usage in 2024?

usecase

I use rlogin everyday in my develeopment LAN environment, it is simple and I can forward the X11 with no drama and no overhead

and I would like to use rcp too, but I can not because of the bug

I also write some tools that might be encapsulated and reusable enough to be interesting to some people out there

so it is not just this particular package

hmm

at rsh

maybe

I have a desktop and a couple of servers with jails, rlogin allows me to move easily from the desktop to the jails

and since it is a controlled environment, you do not need ssh/scp

i'm surprised tho

rsh might have some usages in developing systems too, since you could launch things in another jail

i have used plaintext nfs actually

It is not used anymore, but I am sure other people are using, just they do not mention

locally

funnily was aware of rsh 23 years ago

but ssh was already in use

the r commands are a thing from the 80s

sure

but I think simply people defaults to ssh even in situations where rsh/rlogin/rcp might make more sense

you can also write your own remote services with rsh, it has niche cases

i grew up with ssh to extent that i ssh into local vms

:p

now, that's not bad habit

I think I marginally used rsh/rlogin as a user in the mid 90s in Sun workstations at the university, but it was already a legacy thing

somebody implemented none encryption in openssh tho

for perf

everyone uses rsync

to be honest i've used rsync in windows machine to copy files betweem two local disks

i think i'm too young for uni in mid 90s

if it makes you feel any good Tenkawa, I had a former colleague at work that used to receive the service orders via telex

ice

he worked at Philips, installing x.25 stuff for banks

er Nice

although it might have been good idea to put me somewhere at age of 12

Yeah I was at AT&T/NCR/Lucent

I started on those thogh much younger

I came to industry later, but I still had to learn -and configure at the lab- x.25 stuff at the university

but that was already clearly legacy stuff

uskerine: heheh I hear ya

just nobody mention uucico, y'hear?

haahaa

actually it's to my relief that some of those technologies are gone

somehow

jgh: don't forget fidonet as well

ahhhh, ka9q

r9laj

been using ax25 and fbb

ketas: in an rsync context, rsh is just a useful way to communicate the concept of an rsh-like thing

from an rsync perspective, there's no practical difference between rsh and ssh

it's just the tunnel via which you execute the server-side rsync on the remote machine to conduct the transfer

s/tunnel/conduit/, maybe, to avoid nitpicking

What's the right thing to do when ZFS reports that your block size isn't correct: nda0 ONLINE 0 0 0 block size: 512B configured, 32768B native

last1: "yep, with Docker" Aka, how to distribute your software the wrong way.

crb: i have a few VMs like that; not sure there is anything to do

uskerine: its pretty easy. you are welcome to bug me in #freebsd-ports with questions (or anybody else ofc)

crb: check the block size of the h/w device, if possible I would recreate the pool with the correct blocksize.

atm every 0,5K write will rewrite a full 32K block.

if the blocksize matches, then zfs will be smarter.

imagine writing a 64k chunk to a file, you can either write 2 32K blocks (and the parent blocks also, because CoW)

or 128 0,5K writes that actually each cause a 32K to be written physically on the device

crb: I am interested in the nvme device in use here. you may find some lowlevel parameter that lets you tune this.

kernelnewbies.org/KernelProjects/la…ase_in_block_sizes_on_flash_storage was very informative for me

crb: this is the definitive article openzfs.github.io/openzfs-docs/Perf…0Tuning.html#alignment-shift-ashift but my suspicion is that this nvme device block size is not actually 32k. more common would be 4k or possibly 16k.

crb: a low level nvme format might be possible, I have not needed this in the past so never tried it

crb: the actual command is man.freebsd.org/nvmecontrol

Guten Morgen dch!

yep, nvmecontrol identify on namespace should show the list of LBA formats, and the current one

Oclair: you are German?

when we getting tea

err...wrong buffer, sorry :-|

You should never need to low-level format a block device, unless you're making a change to the block size.

Dealt with this when borrowing Fibre Channel drives from SAN/NAS kit to install on a Windows server.

The SAN/NAS technology used 520 bytes/block, and PCs used 512.

CrtxReavr: was the 512b fs 520b disks solely to piss everyone off or they used ecc there or why?

and did they split a byte and bit for that

There was a technical reason for it. . . I think it was some checksum value or something.

I prefer my 4096 ones personally but anything that uses u-boot and sbi (I do a lot of risc-v and arm stuff) is really annoying to work woth them

s/woth/with