jbo

hmm pkg install alire gets you a huge startup time when you start a new bash, seems the completion is bulky or something

Or another xz

lts: you talking to me?

Yes

lts: eh, I doubt it, its just misconfigured

I was reading the wacom tablet guide on the FreeBSD wiki, it requires webcamd, which is a Linux userspace usb implementation... is the one of FreeBSD a port of this?

"The webcamd daemon is basically a port of Linux USB device drivers into userspace in FreeBSD." ~ webcamd(8)

sorry I should have RTFM

anyone knows if sound card creative AE7 runs on freebsd ?

it seems linux wont even boot with it

but im not sure if freebsd would boot and then would be able to passthru a bhyve with windows

i don't know

how do i set a recording source with mixer again?

i can never find this, i know it's hidden in the man page somewhere… :F

Has anyone managed to get seafile-server working on FreeBSD 14.0?

I've upgraded my system, but seafile-server is no longer in binary packages, and building it manually seems to be broken.

... so I got it to build, but holy shit, the server's codebase is scary bad O_O

(think using T* instead of T** in places [compiler warnings *are* triggered, but appear to be ignored upstream?], mismatched callback/function signatures, passing `char*` as a `_Bool*`, etc)

heavens

This is something I've relied on as a Dropbox alternative for years.

is anyone tried to run FreeBSd on odroid M1 ?

See wiki.freebsd.org/arm and its subpages. (I could only find a mention of Odroid-C1, which is discouraged as a platform due to bitrot.)

Title: arm - FreeBSD Wiki

what're best centralized logging stacks compatible with freebsd? i'm hearing prometheus is good for logs, traces, and metrics

i want collectors on every freebsd host, vm, jail, shipping all that to a centralized timeseries db on another freebsd host

prometheus, collectd, netdata are all good

DaVieS: i've never tried it, but the rk3568 does seem to be supported

ofc that doesn't necessarily mean that particular board works due to vagueries of arm...

realdeimos: netdata runs on freebsd? niiice

i heard collectd is really sluggish and resource heavy

is it better to say "coauthored by" or "obtained from" in commit log when someone gives you words?

concussious: aiui, "Obtained from" is for when you copy and paste a change from somewhere else, so i would use something else. i'm not familiar with 'coauthored by' though

I've been using prometheus+loki+grafana and having good success

I setup promtail on the loki instance to accept syslog and just forwarded my fbsd boxen's logs that way

and as far as collectd, I've never known it to be sluggish or resource heavy unless you're using the exec plugin and calling scripts with a large runtime

concussious: 'reviewed by' might be sufficient unless the other person wrote the bulk of the commit

I use it on my *BSD machines with the output_prometheus plugin and with just a large battery of official plugins loaded, it's using < 10MB of resident memory (that may or may not be heavy depending on your environment I'll grant)

lw: thank you!

i only know this because i put 'Obtained from: Linux' in a recent commit and was told not to do that as it wasn't actually code from Linux

in #1195 I used a sentance imp wrote and also pulled some words from openbsd that make those sentances match exactly

so I said obtained from for both of them

i feel like he went way out of his way to include me in the game, and I don't want it to be like I'm not being respectful

you could put 'Co-authored by: imp; Partly obtained from: OpenBSD' and see if it gets edited before merge, then you know for next time :-)

Thanks!

(take this advice with a grain of salt, i'm not a committer and definitely don't really know what i'm talking about)

not a committer, sure, but I recognize you as much more senior of a developer than I am.

even though your oldest commit in the log is actually the same day as my oldest, I just try and maintain manpages.

i would hesitate to call myself a developer, i just submit PRs for things that annoy me

shh we're backbone software engineers now lmao

though, really, maintaining manpages as just studying the OGs and upstreaming their words.

concussious: you could also use, e.g., "with suggestions from:" or just some free-form text above the trailers crediting him if you'd like

the project is historically not *that* strict on what constitutes a valid trailer

kevans: thank you!

We still can't have multiple postgres servers installed at the same time?

skered: no, somebody needs to care enough to do the work

I guess I could just jail to jail it.

that's what I do... very annoying =)

See wiki.freebsd.org/arm and its subpages. (I could only find a mention of Odroid-C1, which is discouraged as a platform due to bitrot.)

Title: arm - FreeBSD Wiki

OOops.

V_PauAmma_V: I always notice you have really great wording on differential/manpages. Do you have a moment to review a Pr? Also is it impolite for me to ask you?

Title: sysexits: clarify depreciation and history by concussious · Pull Request #1195 · freebsd/freebsd-src · GitHub

I don't find it impolite, personally. Certainly much less so than risking leaving errors or ambiguities for later readers to find. *goes look*

There.

Thank you!

based on that other discussion about exit(2) for usage, i'm wondering about submitting a PR to add some more modern exit codes to sysexits

need to find out how widespread those 'standard' exits are though

realdeimos, OldHead, ever try timescaledb? i like that it's timeseries built atop postgres which is rock solid

realdeimos, OldHead, ever try timescaledb? i like that it's timeseries built atop postgres which is rock solid

lw: not very

lw: in pr above I've s/preferable exit codes/4BSD exit codes/

sorry, s/preferable exit codes/4.4BSD exit codes/

realdeimos, OldHead, ever try timescaledb? i like that it's timeseries built atop postgres which is rock solid

kevans: why can't we run separate pg instances on freebsd?

alepzi: you can, in jails; they try to install to the same paths because we don't do the multiplexing stuff commonly seen on various linucises

Seems like running separate jail containers for each postgresql instance would be the best answer anyway. And relatively easy. So no real itch to scratch to make multiples co-installable natively.

ya i guess that makes sense

rwp: most people don't want to run them side-by-side for long, that's the supported upgrade path

you install two at the same time, then upgrade and remove the old one

so you end up having to run a new jail just to upgrade into it

you can also just pg_dumpall the old version, remove that, install the new version, and restore it

jbo: rather than just go "lw" "jbo" back and forth for days let me just link the thing i was going to show you: bugs.freebsd.org/bugzilla/show_bug.cgi?id=278535 -- iirc you had a problem like this as well

Title: 278535 – bhyve: occasional long delays during boot

lw: good use for the gdb stub to start with, I'd be curious what it's doing (are they all stuck in the loop at the end of init_secondary_tail, or somewhere after that?)

kevans: unfortunately i haven't encountered this often enough to throw any tuits at it but i'll keep that in mind

*nod*

my initial suspicion is there's something wrong with how vmm threads are scheduled on the host, because i've noticed load on the host can often make VMs much slower than they should be

(noticed this with Linux guests too)

probably, though 'several minutes' is kind of crazy bad

maybe unrelated but both systems i noticed this on are AMD

that part of boot is still generally single threaded, I'd expect some progress even if it's obviously many seconds between messages

kevans: yeah, the first time i ran into this i thought the VM had hung completely

it could somehow be an AMD specific thing... I'd kind of expect that others would have seen this by now if it were a larger general problem, but afaik AMD tends to be on the lesser-tested side for bhyve specifically

unfortunately i don't own any Intel CPUs to test on

although i am expecting an Atom device in the post tomorrow, maybe i'll try bhyve on that while i'm setting it up

need to work out if i care enough about ipfw to make it work without INET

that needs an ABI break due to how it works

anyone running timescaledb for observability store?