EuroBSDCon 2026 was my third EuroBSDCon, all in a row, and this year I was determined to finally write a blogpost entry for EuroBSDCon, which I failed to do for the past 2 years, as I left it on the bottom of my list and then when it came around to writing it, the experience was no longer fresh in my memory.
This year will be different, as you can see, because you are reading it :)
This years EuroBSDCon was in Brussels, Belgium. I am from London which has a high speed rail link direct between the two cities, see Eurostar for more information.
EuroBSDCon was also early this year, usually being in the later weeks of September, this year it was in the earlier weeks. I usually take September to meetup with my close friends around EuroBSDCon, usually I do this before the conference, and attend on the way home, but this year was flipped due to the early timing.
Typically, with a direct train from London St Pancras International to Brussels Midi/Zuid, I could get there in 2 hours (3 hours incl the 1 hour time jump), however due to an issue between the chair and keyboard, the booking was actually made to Paris Nord...
Instead of rebooking and forfeiting the money, it was cheaper (and more fun, yes I am a rail nerd) to take a connecting Eurostar (formerly Thalys) from Paris Nord to Brussels Midi/Zuid, with a total travel time between the two trains of 5 hours instead of 2.
Overrall was a fun journey, it had been a few years since I used a Thalys service so it was fun to be on one again. Unfortunately I have no pictures as Paris Nord is filled with firearm officers, and pointing a camera towards them standing in front of the trains felt like a good way to get into trouble.
I have already covered my talk at the Devsummit in a previous blogpost so I will be skipping over it in this section.
Of course I had to wake up late on Day 1, alarm after alarm and still nothing got me up. I was staying 50 mins (walking) from the venue, so I had to leave considerably earlier than most other people attending the conference.
Why not just use the tram? Well I could, but in Brussels you can only top up a mobib card with coin, not cash (or card, but who wants to be tracked, or pay ridiculous transaction fees when abroad? not me!). This makes change valuable in Brussels. I had a few tickets loaded and I didn't want to use them unless I need to, and I also love a morning walk, so I simply walked faster.
I arrived at the venue at about 9:36am, 6 mins past the opening of registration with sweat pouring off me as it was pretty hot that day, so it looked like I had went for a morning run rather than a leisurely walk to the venue.
I was glad to see many familiar faces from last year's Devsummit. This was my second Devsummit, so I knew quite a few people from last year, and was excited to meet new people, more on that later :)
This was a talk I was excited to see. The speaker, Federico Angelilli aka "fedang", had worked on it for GSoC (Google Summer of Code) and although I personally don't support the idea of live patching, it wa still a very interesting talk.
I would later discuss it a lot with fedang, and I am looking forward to potentially seeing a proof of concept of live patched security vulnerabilities.
A quick "why I would personally not use it". I personally feel that for most environments security is more important than a minute of downtime, and rebooting will always be more secure than swapping out functions in memory, which has the potential to go wrong.
A proper infrastructure stack should have redundancy, so the end user sees little to no change in service during security patches, or maintanence. Obviously this is not always possible, but reducing the impact on the customer is better than nothing.
Live patching is vital for certain services though, for example timeservers which are time sensitive (obviously). FreeBSD support for live patching pulls it closer to feature parity with Linux, while also having a suite of better features Linux doesn't support (eg: jails and bhyve within base).
When the recordings go up for the Devsummit I highly recommend this talk, it was my favourite talk at the Devsummit.
As I said above I won't cover my talk here, but I will give a little note. With such a good talk starting off the Devsummit, it was difficult to try to match such a good talk. I hope those who saw my talk enjoyed it, even if I did ramble a little at the end.
Unfortunately for the rest of Day 1 I was stuck in the hallway track. There was little to no plugs in the lecture halls, so I choose to hack in the hallway next to a power socket. I will talk about what I was working on later.
Notably, I got to meet Alexander Ziaee (ziaee@), someone I had interacted a lot with online as he reviewed a few of my doc patches. I did not expect him to be at the conference but I am glad he did manage to attend as it was a pleasure meeting him.
After Day 1 concluded, I was mainly speaking to fedang who tagged along with Benedict to a restaurant which I forgot the cuisine they served. (I am sure someone will shout at me on IRC for being so forgetful :p).
I wasn't really feeling hungry, so I ended up having a few glasses of white wine instead, and simply socialising. I got to meet a lot of the Modirum people that night, and had some very interested conversations.
Funny enough, the other major group of mainly OpenBSD developers (and many of the EBC organisers) ended up at the same restaurant, which I found hilarious.
Day 2 wasn't much different than day 1, got to the venue, drank tea, and then the Devsummit started.
This talk was presented by Jan, and was fascinating. As someone who's worked
with jails a lot, especially if if_epair(4), seeing the benchmarks and
performance gain from stripping out the ethernet frame was interesting.
Along with the higher default MTU, if_pair(4) would allow for much faster
network connectivity between jails than what is already possible with
if_epair(4).
Someone I am definitely excited to pay with in the forseeable future, when I can get my hands on it.
This was a talk presented by Kyle Evans, someone I spoke a lot to on
#freebsd on Libera.chat. He has previously helped
me figure out undocumented loader behaviour, and is currently helping debug
a bug within the FreeBSD loader related to serial. I will make a blogpost on
this in the future, but the TL;DR is Kyle is very helpful on IRC.
This talk was a little too complicated for me, but was still interesting, especially details about nested (hierarchical) jails and startup/shutdown of jails. I will likely rewatch this talk when the recording becomes available as its packed full of information which I would like to digest, and if I can't, I am sure I can ask someone on IRC :)
About this time my laptop was dying again, and I didn't have my spare batteries charged. Back to the hallway track it was.
During this time I worked on improving a patch to fetch(1) with ziaee@, which idiot proofs the documentation to output data fetch downloads to stdout. I wrote it because I was an idiot and completely missed the small mention of it in the manpage, and someone else corrected me. So, why not make an example for it and make it more clear :p
I also continued working on my XMPP patches, at the bottom of this blogpost I will link to all the patches I worked on during EuroBSDCon/FreeBSD Devsummit for those who are interested.
After day 2 concluded, most people were going into central Brussels for food, but I was tired after two days of hacking, and wanted to stay local so I could easily get back to my accomodation, so a small group of us went to an Indian restaurant around the corner.
Here I spoke with Benedict again, but also Ori joined us (more on him later). This is also where I met two Germans which were recommended to come by Henning from the Hamburg BSD User Group, those were Goran and Dominik.
Food was amazing, and there was a lot of interesting discussions as you could expect, but I would return to accomodation after this for some well needed sleep.
For the observant among the attendees, you may have noticed I didn't turn up on the Friday, and there was a few reasons for this.
Firstly, I was exhausted, I wanted a sleep in, as would anyone. EuroBSDCon and the surrounding devsummits and tutorials is a very long conference, and it is difficult to make it through the whole week hacking all day, drinking all night, and then having very little sleep, passing out in talks isn't fun for anyone, nobody wants to hear you snoring.
The other reason I didn't attend (just a few hours into the day) is because of the poor network connectivity and power situation. For the entire conference I was tethering off my phone because the provided WiFi was so terrible that it was difficult to maintain any connection, not to mention it blocked VPNs, SSH and XMPP all of which I need to work.
My airbnb had high speed networking and a kettle, along with plenty of plug sockets to charge both my phone and my laptop as much as I needed. It seemed pointless to travel to the venue to sit around waiting for things to load.
So instead I walked down the road, bought some snacks, and a box of teabags, only Lipton Yellow Label, which if you are British you know that this is far from good tea, and then walked back.
I must have drank 15 cups of tea that day, and for those who are wondering, no I do not pollute my tea with milk. Oh and also, tea requires hot water, the water at the venue for the tea was only warm, which led to the tea tasting watery instead of properly brewing, which was a shame.
Nothing really more to this day, I relaxed and hacked.
Finally getting to the main conference, the bit everyone is interested in, not me rambling about tea.
I was excited for this day the entire conference, because the talk I wanted
to see most was Stefano's talk. For those who are wondering how I keep track
of talks I want at conferences, I download the ical file for the talk, and
load it into the calender on my phone before turning up at the conference,
any blanks in the timetable is "hacking time".
I can't comment on Hansteen's keynotes, because I will be honest, I couldn't hear him at all the entire time. This was a common issue in the audience too from what I heard.
The talk I was most excited for, and it delivered.
Stefano is amazing at mixing humour, which genuinely good advice, you leave his talks not only having enjoyed yourself, but also a lot wiser than before the talk.
The TL;DR of the talk is always have backups, and not within the same physical location. It sounds like common sense, but Stefano wisely pointed out that in practice many individuals and businesses do not.
I highly recommend watching the talk if you haven't, more information here.
This talk was presented by Lukas Engelhardt, someone I do not believe I have interacted with yet. Nevertheless it was a great talk.
I will be honest, the main reason I attended the talk was to try to convince me that pkgbase was actually a good thing, and I left still unconvinced. For those who want to learn more about pkgbase though, it is an amazing talk, check it out here.
I would later speak to both ziaee@ and rene about pkgbase, where I would somewhat be convinced about it. Maybe I will do a blogpost at some point explaining why I dislike the idea of packaging the base system.
In any case, I did think of a "solution" to my concerns about pkgbase, in other words, how to script pkg to behave how I want it to, this I will document when I have the time to script it.
This talk was presented by Martin Vahlensieck, another person I have not interacted with personally.
The main motivation to attend this talk was that I ran into the exact same issue but on FreeBSD, so I was interested to see the OpenBSD way of doing it, seen as I also use OpenBSD.
For those who want a rough comparison, OpenBSD has routing domains to separate routing tables, while FreeBSD has fib (see setfib(1)).
Quite a niche talk in my opinion, but definitely worth watching if you use Wireguard on OpenBSD. More information can be found here.
This talk was presented by Charalampos Mainas, someone else I have not had the chance to interact with in person, and in the usual trend with EuroBSDCon, it was also a great talk.
My motivation for attending this talk was due to a non-profit project I work for whos stack rely on containerisation. I thought it would be a cool talk to attend to then potentially try to run our software within FreeBSD containers, on a Linux host.
Why? Well FreeBSD is much smaller than a lot of the Linux base images, not just that, but the security of FreeBSD is much easier to track as its one base system, one src tree. Linux on the other hand is a bunch of different projects glued together, therefore when you have a vulnerability within your init system, this is updated independently to a vulnerability in your crypto library.
This makes FreeBSD an amazing base image for containerisation, and this talk did provide the motivation for me to look into this more in the coming year. More information can be found here.
Some miscellaneous stuff which people might find interesting.
Firstly, I dumped some XMPP stickers on the registration table and on the tshirt table along with the other stickers there, surprisingly these went really quickly. I did not anticipate them being popular. Will need to bring more next year >:)
Secondly, the Modirum folks brought a few kegs of beer with them, and had cracked one open at the back of the hallway near their stand. I was waved over, and offered a glass of beer, which I happily accepted. The beer was from Kiel, and had a really nice flavour to it.
I also had a very interested discussion on how annoyingly useless AI can be. Thank you for the free beer, and the great conversation!
Finally, the artwork and the tshirt this year was amazing, I think this is my favourite tshirt so far out of 2024, 2025 and 2026. The artwork is also on a dark-greyish black tshirt, which stands out to the other two I have, which are both white with the artwork printed on it.
The long anticipated social event at the Africa Museum was finally here.
Unfortunately the tram wasn't running due to engineering work so we had a 20-30 min walk down the road to Montgomery where the organisers had laid on 3 old trams to bring us to the museum.
As someone who loves public transport, when I heard that there will be old trams bringing us to the museum I was really excited, and I was not disappointed.
Below you can find some pictures I took:
And for the fellow public transport nerds, yes I have videos of the tram departing, onboard looking outside and onboard looking inside, enjoy!
When we arrived, the size of the museum was amazing. No wonder it was so far from the city centre when it sprawled, with tons of green space, and beautiful buildings.
When entering through the front, we had to get our badges scanned, and after passing through and going upstairs, I was shocked by just how formal it was. I am probably one of the least formal people you will meet in life, and I definitely felt out of place, we were given champagne on entering, which by the way, tasted really good.
We were then encouraged to go on the tour while the food was being prepared.
Honestly there was a lot of people, and limited staff to do the tour, so in the end me and a few others ended up walking around on our own looking at stuff, but the Museum being so big, and with us only have a few hours at the Museum before it closes, I ended up skipping through large portions, unable to read a mere fraction of the information available due to limited time.
I was also exhausted from the long day, so I didn't really have the energy to read, which was the disappointing thing, a Museum is best enjoyed with a lot of energy, and we were there during the night when most people were tired, and probably didn't have the energy after a day of hacking to take in the information.
Nevertheless, I still enjoyed walking around the museum.
Upon coming back up to the main room, we were greeted with a buffet of food, from BBQ chicken, to chips (fries for you Americans) to fresh salad and bread, there was so many choices, I honestly can't remember them.
Quite early on I realised we had 2 hours to eat and drink before closing time, and the only beer available was stella, and was being poured into 300ml glasses, and because it was coming up poorly, half of it was a head, so you were effectively drinking 150ml per round, which was impossible to get enough alcohol to even feel the slightest bit intoxicated.
So me, Goran and Dominik decided that the wine was a better choice, we would go up and order 3-4 glasses of white wine each every 10-15 mins, and ended up lining up glasses and glasses of white wine. We filled the table numerous times with wine glasses to the point it was almost embarrassing when the staff came over to collect them all.
Closer to closing time, we ended up having plates and plates of chips to counteract the munchies from the alcohol, and them being salty then made us want to drink more, and by the end of the night when they stopped serving, we were the last group of people who still had alcohol leftover.
It was a great night, and I got my money worth from the social event ticket, but still left the place not feeling that intoxciated, just chatty. So we decided to go with Tom back into Brussels for more beer, in which we stayed until about 3am before parting ways, me getting back to my accomodation at 4am.
Overrall, I was worried about the social that there would be limited drink as its a museum, and that the food would be meh, but I was shocked by just how great the service, the assortment of alcohol, and the food was. I have only been to two social events, Zagreb and then this one, last years was definitely a bigger p*ss up than this year, but this year had the better food, drink and atmosphere overall. The only improvement we could have had was more time to hang out, there were people I wanted to speak to and we only were just getting started when the night had to come to the end, which was unfortunate.
So on day 2 there wasn't much I was interested in the morning, so I planned to just take it easy and hack, especially with the amount of drink the night before.
The first talk I wanted to see was on GEFS by Ori, another talk I had been anticipating the entire conference. However, idiot me forgot to set an alarm the night before, ended up snoozing, and because the trams were down for engineering work, I couldn't get there any quicker than speed walking. I did however, make half of the first talk I wanted to see, which was better than nothing.
As I said above, this talk was presented by Ori Bernstein, an OpenBSD and 9Front developer, and generally a very interesting person to talk to.
The motivation to attend this talk was, well I knew Ori, first meeting them at FOSDEM 2026 after my talk on OpenBSD networking, which wasn't the first time I was introduced to GEFS, which was before when Ori first suggested he would port it from 9Front over to OpenBSD, which a friend of mine told me about.
I was sceptical of the filesystem at first, and of the OpenBSD team ever allowing it to be merged into the src tree, however the closer Ori gets to completion the more and more likely it seems that OpenBSD will finally have a proper filesystem, shocking I know!
Personally I am interested in GEFS from a networking point of view, because currently OpenBSD based routers which get unplugged by the user, or are subject to a hard poweroff due to other factors (such as a power cut) can result in filesystem corruption, usably fixable with fsck but sometimes, permanent.
As you can imagine, having your router spontaneously corrupt due to accidentally unplugging it before fully powering it off gracefully when moving it (as an example) is not a very good user experience, and is horrific from an enterprise point of view.
GEFS should hopefully bridge this gap, and finally solve this issue, where others I know simply script fsck to discard any errors on boot, which is far from a decent solution to the problem.
Further information can be found here.
This talk was presented by Dave Cottlehuber (dch@), someone who has done a wide range of talks at EuroBSDCon in the past, and a very interesting person to talk to, and one of the few you can contact on IRC too!
This talk was of interest due to the increasing concerns about security updates lagging behind on FreeBSD. As dch@ showed within the talk, there has been a massive surge in vulnerabilities which the FreeBSD security team has had to handle.
It would be interesting if someone looked into the same thing but this time for ports, as major ports such as xorg-server (which runs with SUID) often lag behind on updates, which is a major concern for those who use FreeBSD on the desktop.
Further information on the talk can be found here
This talk was presented by VinÃcius Z, someone I have not interacted with before personally.
The main reason I attended this talk was because I was planning to attempt something similar, after spitballing with Kyle Evans on IRC. I wanted to see what he had done, and was a little disappointed to find it was more of a proof of concept rather than a fully working solution.
I will have to review the talk again as it was hard to hear him during the talk, and then at some point in the next year I want to attempt my idea which is quite similar to his proof of concept to see how well I can get this to work.
More information on the talk can be found here.
This talk was presented by Charlie Li, someone I have interacted with quite a bit, and originally met randomly in a bar in Zagreb during EuroBSDCon 2025.
This was another talk I was anticipating, but during the talk I was feeling really unwell so I had to take a step outside and I missed a bit of the talk which was unfortunate.
I have worked on a few python ports, so this was of interest to learn more about python porting. I will likely rewatch this at some point to fill in the gaps I missed. If you are into porting, or planning to port a python port, I definitely recommend checking out the recording for this talk.
For more information, see here.
This talk was presented by Pierre Emeriaud and Thomas Pasqualini neither of which I have interacted with myself.
This was the final talk of the conference, and I was drawn to it having previously done networking on OpenBSD for my own server network, so I was curious how Orange runs their network on FreeBSD.
Furthermore they dive into making images using poudriere-image something I
have not played with before.
Definitely gave me some inspiration, and I might try deploying some FreeBSD-based routers around at some point, which tend to be more performant than OpenBSD, although not nearly as fun to configure and as small in size. The simplicity of OpenBSD, especially when it comes to network applications, is a really big perk which is hard to compete with.
For more information on the talk, see here.
The closing is always a fun experience, with Henning's humour, announcements on other BSD events (BSDNL, AsiaBSDCon, BSDCan), the auction, and best of all, finding out where the next EuroBSDCon will be!
I won two items on auction, a signed copy of "The Book of PF", I already have one which I bought during EuroBSDCon 2024 directly from Hansteen, however that was the 3rd revision, and this was the new 4th revision which had been much anticipated. I was surprised people didn't bid as high for it, and I got it for 60EUR, also with a signature from Henning. I also won a small bottle of fizz for 5 EUR which I was also surprised nobody went for, their loss I guess :)
I would later drink the fizz during the social after the end of the conference, it was a decent bottle of fizz, but unfortunately, warm. The PF book went to my friend who really wanted a copy, and I was willing to hold off until I could pick up another copy.
The next EuroBSDCon will be in Bratislava, Slovakia. Dates are still currently unknown but you can check the EuroBSDCon 2027 Website to stay updated on next year.
So the original idea was to meet up at the speaker hotel and go somewhere else, but in the end people gathered in the speaker hotel and drank there, made it easy for most to go to bed after a few drinks, and for others, get rat-a*sed.
Pizza and beer, what more could you want in life?
In my case, I was a little antisocial and was working on my patches, which fedang and Charlie Li stayed up to 6am hanging out, and in the case of Charlie, helping me with a few nuances I couldn't figure out.
I also did socialise inbetween working on the patches, I was desperate to get them submitted by the end of the conference, but by 6am the following morning, when people were getting up to checkout of the hotel to catch an early morning flight, all port patches were finished, and ready to submit, I succeeded in my goal, but at a major cost to my sleep.
It was a pleasure seeing you all again this year, thank you all for the great time, and to the organisers of this year.
Special thanks has to go to:
See you all next year, or else!!!