Hardcode
ECMP is one of those
i686
ecual cost multipath is one of the worst protocols that can be used for load balancing
it will work, but it is not even a good idea
Anonymous
You can git clone/compile any program you want, if its meant for BSD or the depenencies can be found in a FreeBSD system. You'll need to tweak where the binary goes, or the Makefile to point to /usr/local deps and such to have it working
does it still work?
> https://riptutorial.com/freebsd/example/23758/packaging
ɴꙩᴍᴀᴅ
does it still work?
> https://riptutorial.com/freebsd/example/23758/packaging
Dunno, never used it. Best thing to guess it, is by trying it. I suppose it should
Anonymous
https://archive.org/search.php?query=freebsd&page=3
Hardcode
does it still work?
> https://riptutorial.com/freebsd/example/23758/packaging
this works only for ports. means that it won't work for anything that is not it ports. in fact, this article just briefly copypastes chosen parts of the official documentation. it's useless when you have something that it's not in ports. there's a FreeBSD Porters Handbook, which thoroughly describes anything you need to make a port, but if I were you, and just needed a one-time software that I don't have plans to distribute across dozens of servers, I'd just compile it into a custom local prefix. i.e. ./configure —prefix=/usr/local/mycoolsoftware-0.0.0.1 && make all install (if it's using the autotools, which is still by default a decent choice)
Hardcode
in fact, "building into a local prefix" is a unviversal way to get just anything on any UNIX system, not just FreeBSD. there are tonns of software that are distributed in source or missing the binary artifacts for your type of Linux/Solaris/whatever
Anonymous
Cool
Hardcode
there can be complications though when it's not assembling on your system because of missing library/header, or worse cases, but that's a whole another story. 90% of software can be easily built.
Anonymous
Anonymous
I hope that this great OS live as long as possible =)
Hardcode
not quite. usually all you need can be found in ports.
Hardcode
I hope that this great OS live as long as possible =)
I doubt it will outlive Linus. I think as soon as he dies, Linux will finally get the decent zfs support (not the one it gets now) and the FreeBSD user base will diminish by a certain multiplier
Anonymous
Hardcode
looks like Linus is still alive and kicking, so nope.
Anonymous
looks like Linus is still alive and kicking, so nope.
Well, as long as it's open source I think that it will lives
Anonymous
Freebsd seems very portable
Hardcode
liveness is a discussible matter. is Illumos alive ? is Openindiana ? Some think it is. I really doubt. Well, formally it's alive as long as there's at least one developer, but that kind of life is somewhat creepy.
Anonymous
Hardcode
yeah
Hardcode
FreeBSD now uses the same upstream and many ppl think "okay it's the same one", but that's not quite right. FreeBSD got way more advanced zfs kernel support, and routine zfs operations in Linux - like expanding the pool or replacing a disk in a bootable pool are pain in the ass. in FreeBSD or Solaris those are boring and predictable. in Linux those are a terrible and horrifying quest
Hardcode
when a disk dies in a bootable root pool, Linux often is even unable to boot automatically after disk replacement (and not because it doesn't see the loader, but because it refuses to import the pool).
Hardcode
Linus is effectively blocking all the kernel zfs support
Anonymous
Oh, so we are waiting his death
Hardcode
because he has religious contradictions with CDDL licence the zfs came under
Anonymous
because he has religious contradictions with CDDL licence the zfs came under
Well, seems like our savior is that, the license
Hardcode
Linux can use zfs for sure, many use it even in production, but this is a life full of adventures.
Anonymous
Well, seems like our savior is that, the license
Without it, maybe Linus would accept implement the feature and freebsd would slowly die
Hardcode
and that's why Free/TrueNAS still struggling with migrating to Linux i guess
Hardcode
they announced that there will be Linux instead of FreeBSD long time ago
Hardcode
Oh, so we are waiting his death
well, I think in that case Linux will dissolve into many proprietary distrubutions, not quite compatible with each other.
Anonymous
Hardcode
so - no, without Linus things will only get more interesting
Anonymous
so - no, without Linus things will only get more interesting
But I don't wanna freebsd to die
Anonymous
Anonymous
so - no, without Linus things will only get more interesting
Imagine another red hat like, but with applications that are incompatible with other Linux
Chris
Linus is effectively blocking all the kernel zfs support
is that true for the DKMS kernel modules of openzfs too? I use zfs on linux but only for basic stuff on a desktop machine with single harddisks though
Hardcode
Hardcode
zfs is FreeBSD fs of choice, unless you're on some embedded stuff where you still use ufs. and on Linux zfs is for enthusiasts only. you need full support from kernel - whether it's ARC and virtual memory interaction, or kernel-level iSCSI target, or exsessive TRIM operations
Hardcode
Linux has gotten some basic zfs support, but there's still Linus who says "we don't need it" meaning "I don't need it", so it's still very basic.
Chris
perhaps Linus can tell me where I get a filesystem with native encryption, without putting it in a LUKS container that I cannot open with another OS outside of Linux
K
I hate stacking lvm on top of luks on top of a filesystem. I just use ZFS who has everything
Chris
But I don't wanna freebsd to die
iirc Netflix uses FreeBSD so there will be always use cases, don't worry, FreeBSD has not died since the past 15 years of "this is the year of the linux desktop!!!!!" :D
Chris
just look at the distro graveyard on distrowatch, distros come and go but all BSDs have been consistent and stable like forever
Jay
perhaps Linus can tell me where I get a filesystem with native encryption, without putting it in a LUKS container that I cannot open with another OS outside of Linux
You can open LUKS container with Windows and DragonflyBSD
Chris
interesting, if I choose to install windows ever again this life
Anonymous
This game is awesome
https://www.freshports.org/games/openmw/
Anonymous
XS
in fact, "building into a local prefix" is a unviversal way to get just anything on any UNIX system, not just FreeBSD. there are tonns of software that are distributed in source or missing the binary artifacts for your type of Linux/Solaris/whatever
Totally true, I'm also doing it a lot at work for many platforms where we use a arch/software/version/ schemas for prefixes. It's really flexible.
Anonymous
Totally true, I'm also doing it a lot at work for many platforms where we use a arch/software/version/ schemas for prefixes. It's really flexible.
This is cool, I didn't knew that was so easy to create a isolated environment like that
Anonymous
I always thought that we have to do a "make install" and it's done =P
XS
in fact any decent released source code has it's own configuration, build and installation structure that most universally allows you to set the prefix you want and many others parameters regarding features and dependencies that you have the whole freedom to tweak
XS
this way you can "easily" run multiples versions of softwares, depending on multiples versions of libraries, together
XS
that's definetely not a recommanded way of managing a workstation or a production server I would say
XS
but in our cases it's really useful for multiplatform QA and portage purposes
Anonymous
I always thought that we have to do a "make install" and it's done =P
since i'm a newbie on this way of compiling, i cannot say more than that
XS
it require some specific knowledge and can lead to unstable results if it's not done carefully
Anonymous
i always used the pm of gentoo (2y and counting) and it manages all, i really want to try freebsd because it really have a nice history and the implementation is gold (a real unix born)
XS
but at the end, it's not that complicated to do that stuff manually
XS
it's just eating a lot of time
XS
hmm in slackware you have provided source and binary ports ins't it ?
XS
basically you can use any unix system, installing a cross platform port and packages ecosystem that avoids you to do manual builds, you just triggers them but unfortunately, you don't have much control on target installation prefixes
XS
and less flexibilities regarding running multiple independent and conflicting versions
XS
on the other hand, doing everything manually requires you to know and build exactly what softwares you want does require before being able to build and run it
XS
this can take some considerable time getting them right, compared to a ready to use port-tree
XS
you'll also need to track and build updates manually
XS
in my opition port systems are great and real time savers, and provide robust results for most situations
Anonymous
bittin-
Listening to todays BSD now on my way to the dentist will join the online FreeBSD DevSummit later today and tommorow too
ɴꙩᴍᴀᴅ
https://wiki.freebsd.org/DevSummit/202206
bittin-
👍
bittin-