pfSense® Software Embraces Change: A Strategic Migration to the Linux Kernel

planedrop

@AndyRH I too want stability, though I think it's worth saying that updates should be done anytime they are out regardless.

either way, I think with Linux being used we could see some nice features and possible better performance.

AndyRH

I actually update within a week of availability. I have to give my users notice, without notice the wife and kids can be mean.

SteveITS

@AndyRH The part I'd prefer to avoid is

https://docs.netgate.com/tnsr/en/latest/releases/release-notes-22.02.html#operating-system-change

planedrop

@AndyRH Yeah I get that for sure, didn't think you were implying to just leave things un-updated, just always like to note it in case people find threads on Google way later; only because so many seem to think updates are still not necessary looking at how many unpatched things exist lol.

AndyRH

@SteveITS I was thinking about that. By far not an expert in ZFS, but would it be possible to install the new OS into a 2nd boot environment? I suspect this would not work, but it would be cool.

johnpoz

You guys do understand this is April Fools joke right? heheh

AndyRH

@johnpoz said in pfSense Software Embraces Change: A Strategic Migration to the Linux Kernel:

heheh

I think there is a very high probability you are correct. But a port to run on Power HW would be a speed game changer...

planedrop

@johnpoz IDK, is it? I think it's not, most think it's not.

It feels very legit, Netgate hasn't confirmed anything about it online and it seems like a lot of effort to go through for something that isn't really funny.

TrueNAS did something similar saying they are moving to a Windows App (lol) but it is very obvious it's fake, not just because of course not, but it's written in a very silly and sarcastic manner and is a forum post not an official blog post.

It also said at the bottom that it is an april fools joke, this does not.

johnpoz

@planedrop haha - see the other thread about this where Jim himself Chimes in pointing to their other joke from 2017..

https://www.netgate.com/blog/building-a-behemoth-router

https://forum.netgate.com/post/1160479

Nobody is going to release such a statement on 4/1 without it being meant as a joke..

planedrop

@johnpoz OK sure, but they put this at the bottom:

But yeah I see the link, guess they got me good.

LarryFahnoe

So now that we're past April 1, my comment is that the 2018 announcement portended TNSR, though it certainly also made humorous references. I don't know Jim Thompson, but I do know that having something of a light heart is often a benefit in business. So does this year's announcement simply foretell another chapter in Netgate's evolution? It would seem logical to build upon the work that went into TNSR and the experience gained to create a new firewall product. Certainly seems logical to capitalize on the larger/broader development community, though as the recent xz issue illustrates, targeting and dominance are related. Rather than bemoaning an announcement like this, I see the opportunity to bring the experience gained in the FreeBSD world to bear upon tools from the Linux world. Design skill and methodology are often as or more important than the tools themselves.

--Larry

serbus

@SteveITS said in pfSense Software Embraces Change: A Strategic Migration to the Linux Kernel:

...and from the Reddit link I posted earlier:
"The FreeBSD userland (and musl for libc) is actually true"

...as we watch senior netgate engineering talent wander the aisles of the chimera linux repo...hmmm...

John

grahamperrin

@keyser said in pfSense Software Embraces Change: A Strategic Migration to the Linux Kernel:

frankenmonster

No, and the dates are all wrong.

Not this month; I'm reliably informed that frankenSense and myrrhgration will be officially announced on Christmas eve.

jwt

@Gertjan said in pfSense Software Embraces Change: A Strategic Migration to the Linux Kernel:

@CreationGuy

Thanks for making sure that I wasn't the only one reading the blog - RSS feed.

Serious ...

Migrating to .... a Linux kernel with a FreeBSD userland.

You can stop reading after that phrase. I'm not sure what they are smoking over there in Texas, but I want some too.

Oddly, I was in Colorado when I wrote that blog post.

This :

Following in-depth evaluation and collaborative discussions ...

Where ? With who ?
For me, this is a one man's discussion with the one he sees in the mirror.

I assure you there are more than two people in the conversations.

Or, and why not, I totally do not understand how this might be possible.
( without a whole lot more work - entering new uncharted territory and probably sinking the boat called pfSense with the captain called "Netgate" )

Prediction of failure, Film at 11

Btw : that's a hint : no one dared to put its name at the top.

My ego is boundless.

jwt

@johnpoz said in pfSense Software Embraces Change: A Strategic Migration to the Linux Kernel:

Nobody is going to release such a statement on 4/1 without it being meant as a joke..

My name is Nobody

Nobody

(I mean, Franco did once call me the “pfSense Cowboy” and I am originally from Nevada, so it fits, right?)

jwt

@grahamperrin said in pfSense Software Embraces Change: A Strategic Migration to the Linux Kernel:

Not this month; I'm reliably informed that frankenSense and myrrhgration will be officially announced on Christmas eve.

francoSense already exists.

kstrider

if this is the case it may help out PPPoE cusomers

jwt

@kstrider indeed, but we’re also going to fix the situation on FreeBSD with a new if_pppoe driver that doesn’t use netgraph.

Netgraph is the source of the single-threading and resultant performance issues in FreeBSD’s PPPoE implementation.

Not wanting more dependence on netgraph and it’s architectural flaws which directly lead to severe performance issues is also the reason we chose to not implement netflow reporting like the other project did, and instead wait to provide it in a performant and more accurate implementation, directly in pf.

Eliminating netgraph was also the reason we did the work to eliminate the need for a netgraph module for AT&T fiber customers who need the modem to be on-line to respond to EAPOL frames.

PPPoE is the last subsystem in pfsense that uses netgraph, and active development is underway.

Thus, the long-standing complaints about PPPoE performance will be addressed, with new work from Netgate.

kstrider

@jwt ohh thats good to hear is there any sort of estimated timeframe for this change.

slu

@kstrider said in pfSense Software Embraces Change: A Strategic Migration to the Linux Kernel:

@jwt ohh thats good to hear is there any sort of estimated timeframe for this change.

Yes I guess 01.04.2025, but not sure...