I got back to some of my roots, but optimized. I really like bash, and so I found ble.sh to be quite nice, but not just that. I needed something that could kind of put the glue together and let me work on my own additions as well. I would up with the idea of using bash-it for bash to manage my own plugins and work with existing stuff, like a plugin/modular way.
And while I do like bash, it can be a little less pleasing to look at in some ways, but I wanted some level of consistency too. I ended up also returning back to zsh with zim, but the problems that drove me from zimfw, the tab-completion on an `e` character seemingly randomly repeating itself when I hit that ever glorious TAB key to complete a command or argument, no longer happens, and my zsh experience is amazing.
So what was next? I needed a prompt that I appreciated, and would work for me, and more so, would work on everything I used, preferably. I found liquid-prompt, the only viable solution that worked with both Zsh and Bash and allowed me to make my own custom stuff, even my own custom theme style in a way that I wanted it to be.
Now, I'm primarily back to zsh with zimfw, many of my own plugins I've written and designed for it in my own git server, and I'm once again a happy camper.
I'm also still using chezmoi. So if you'd like to check out what I've setup, you can check out my dotfiles config [on my forgejo](https://git.linux-help.org/psi-jack/chezmoi-new).
Enjoy!
-- Psi-Jack
]]>
For whatever reason I’m getting few issues (not no issues mind you) that are easy to get around even needed. One such issue was a actually with Firestorm, a GUI client for Second Life that I use often, stopped working after updates to Mesa. With the power of DistroBox, and distrobox-assemble, I simply made a Ubuntu 24.04 “firestorm” box that could easily be recreated and utilized it to create my stable functional environment. Using the equivalent of a container as “an application” that just works. I’m actually quite proud of this!
I’m running EndeavourOS with the whole KDE Plasma on Wayland experience. I’ve got my desktop setup one way without hibernation. My tabtop setup with encryption, but still fully functional hibernation, encrypted swap and hibernation file. It’s all been rubbing rock solid. I’m using a mix oh chaotic-aur and paru building the test that’s not in chaotic-aur. Overall I’m thinking this is a very solid setup.
– Psi-Jack
]]>So I've been doing a lot of tinkering in the recent months, hence a bit of the quiet time of not blogging here. I've been playing with a couple distributions I've never really tried before. Specifically Void Linux, which is a rolling-release Linux distribution that's quite actually more impressive than I anticipated, Siduction, which is basically Debian SID with a Calamares GUI installation and some nice customizations the Siduction team's done, and both of these distros are arguably both pretty awesome.
Along with this, I've also been tinkering with alternative operating system options for the Raspberry Pi, instead of the defacto standard Raspberry Pi OS (Formerly Rasbian). With that I've tried Arch Linux ARM, Void Linux as well, and Debian Raspi Trixie.
Void Linux is something I'm impressed with. I don't entirely like it, but I like the very basis of how it runs, and how fast it is for some of it's tooling. xbps-install, xbps-query, xbps-remove, are annoying tools that are different rather than being rolled into one package-manager and/or repository manager as many other distros do. But it's very fast. Using zstd compression which is proven to be amazingly fast, as shown on Arch Linux's pacman.
I even managed to make my own XBPS package from the xbps-src repository. It's similar in many respects to the way Arch Linux works with PKGBUILD, but has some differences for sure. Some of it is even automated to the point you may not need much in the way of actual instructions to run. Just metadata about what it's called, and description of it. Just depends on what's required to build it.
This is a very cool Debian-based distribution from the belly of SID, the unstable branch of Debian, which is still more stable than most rolling-release distros in many aspects. Yes, you can be bitten by it, just as with any distro, but usually in different ways, not because of bleeding edge software because even though it's SID, it's not always 1:1 newer versions of software like Arch Linux and Void Linux would be.
If EndeavourOS and/or Arch Linux ever becomes a burden to continue to maintain, Siduction's the most likely distribution I would switch to. Just my biggest concern is how they still seem to be a version old on KDE Frameworks, which caused many issues on Debian 12, by the very same thing KDE Frameworks being 1 version older than current. But, I think that'll change once Trixie's out.
As I've been poking around at new distros, I've also been trying out new distros on my Raspberry Pi's!
I used to use these in home automation, but since moving out of the country and into an apartment, I'm no longer with a home automation setup anymore, at least for the time being. But, I want to try to use these for something.
Currently I'm running the RPI-4B with Void Linux with the KDE Plasma 6.3.3 desktop environment, mostly to run a browser, with Deskflow to access from my desktop. I've considered using it for just that, a browser for notes, research, game walkthroughs, etc.. But it's sluggish for sure, and accessing it can be tedious in it's current design. I'd like to make use of all these RPI's for something. I've had lots of random ideas.. Barring kubernetes (I will not do this again!), I'd like ideas! Any ideas? Leave some comments down below!
I've tinkered with Void Linux, which is actually amazingly fast, to maintain, to update, to install. I almost wanted to use this, but at the same time, I feel it has some issues that may cause some functionality, but we'll see.
Debian Raspi Trixie is what I have on one of the RPI-3B's now. And it's slooooow to boot. Slooooow to update... I wanted to try this because newer software than on Raspberry Pi OS and all, but, yeah, it's likely going back to Raspberry Pi OS, or Void Linux depending on what I end up using it for.
]]>
So, back in 1997, I started using a desktop environment known as KDE, or at the time actually was known as the Kool Desktop Environment. And it was, it was pretty nice, cool, and was basically revolutionary. Over the years, 1.x came out, 2.x came out, 3.x was amazing, but starting to hit some interoperability snags between other environments, most notably GNOME. 3.x was very solid, however, but things were definitely making changes to some different ways.
4.0 was the first entirely new build of KDE, and the first start of Plasma, the new desktop environment itself, spinning off from their origins to somehow rebrand themselves so to speak. Long story short, 4.0 was actually the worst release in the history of KDE’s DE to date, because it seemed like it was the most unstable, untested, unrefined thing they ever built. It took major bug fixing to bring it up to the stable DE it had been since 3.x.
5.0 later came out, and somewhat redeemed themselves, but not entirely. Functionality that I loved from 4.x was ripped out of 5.x without any rhyme or reason, and new, unusual things were added. Ripped out was the ability that I loved, allowing you to attach multiple windows to a single frame and turn the titlebars into literal tabs to switch between. kwin rules even existed to make this automatic. I loved it, but in 5.x it was gone, like it never existed. What was added was this new concept of “Activities”, which was similar, but different, somewhat, to Virtual Desktops. They both basically did the same thing… But not entirely. It became confusing. Why have two ways to do the same thing? Acitivities compounded the issue more and more by having multiple Virtual Desktops per Activity….
Enough of this! I said, and for a while, I moved to GNOME after, surprisingly, someone form GNOME’s PR team reached out to me to talk to me, and give me a point of view to consider and try with the new GNOME 3.x. I did. And… Long story short, I used it for many years.
There’s so much to say about GNOME, but this isn’t so much a story about GNOME as it is about the era of changes, and how things flowed in my past. GNOME is pretty solid, it’s UX has always been very nice, and accurate. Native GNOME apps just worked, and generally worked well…. Until Wayland started to become the popular thing. New bugs in GNOME started to work, for me still in X11, and no matter how many times I went to the GNOME developers to report bugs, their first and literally only response was “Does this happen on Wayland?” Not caring at all about X11 anymore. Their goal was Wayland. And by all means, Wayland is pretty awesome, for what it is.
Wayland… However, is being rushed so fast, so furiously upon us that there’s not even all the useful tools we older users of Linux have come to expect to have and use. There’s some equivalents for some things, but there’s currently no wmctl, Input-Leap (Fork of Barrier, Fork of Synergy) is coming along, but needs various support from different DE developers. GNOME implemented support, but claimed it was “Supported” before it was even ready by the actual Input-Leap development side, because honestly, how can you really develop and test something that has no support for what you need, until after it’s actually supported and released? Chicken and Egg, GNOME announced too early thus making it technically a lie.
Anyway, getting ahead of myself here. GNOME is solid, great, but has issues. It’s strictly CSD-based (Client Side Decorators), which can be nice, but when you look at the variety of applications that don’t even look the same because they have their own CSD styles and such…. It starts to get a bit old, and the lack of conformity grows larger from the norm. Then extensions, theming, a lot of things that GNOME is literally pushing off. Completely changing extensions from one JS format to another breaking all past ones. Breaking every extension every single release. Theming is similar but even more blunt, Adwaita is literally going the path to NOT be themable, to make it “more stable.”
Well, hmm… This seems a very direct path to consistency, while hindering customizability. It’s an interesting approach, dumbing down.
So, KDE Plasma 6 came out in the Arch Linux official stable repos, and I’d already fully switched off to Xfce with a very unique setup of systemd, and autorander methods to automate some factors of it. But, Plasma 6 was finally something that was solid. Is it Qt6 that’d already been well polished? I’m sure that had some to do with it. But they did things that were interesting. Things worked better, smoother. Still some bugs to work out, already on Plasma 6.0.3, but it’s not nearly as bad as the 4.0 release or the 5.0 release either. This was actually reasonable on the first release.
Still, they have some things that just make me cringe, like Activities is still there, and I’d thought it was on the chopping block…. Apparently someone in KDE’s team has to have it, though! The Dashboard, though, a much improved touch that I just really love.
A couple things really helped me consider going back to KDE Plasma though, beyond the new release. The native support (In Wayland at least) for VRR. The fact most of the applications I use are Qt-based, Electron based, or GTK-based but not using the system libraries for those (Thunderbird, for example). Pretty much the majority of software I was using was actually more dominantly Qt-based than GTK based which kind of helped influence me a little (more like a lot).
So, for now, KDE Plasma is back on my primary day-to-day Desktop Environment, and hopefully this remains a good thing.
Oh! I did say I dropped EndeavourOS for Arch Linux actual? Yes, yes I sure did! Most of this is because, while EndeavourOS is good, they still live primarily on Telegram, and I just can’t support Telegram as a previous blog of mine mentions. Telegram is literally the worst social media platform of them all, and I stand by that statement fully.
But it’s not just that. archinstall has grown, and has become quite a thing of beauty actually. I was generating some form of nightly custom-install thing for EndeavourOS’s customizable installer, so in the very rare scenario I would need to reinstall my system to recover from a major catastrophe, I had a nice means do to so. Well… archinstall’s JSON file method, is so much easier and better to work with, that it makes more sense to simply use that, and integrate that into my restore process automation. So I did. Now I have a higher level piece of mind, nightly backups, snapshots, and a near fully automated recovery process that can get me back up and running in 20 minutes or less.
So, that’s my story for now. Hope you enjoyed!
– Psi-Jack
]]>In trying to get all this working, I’ve been running into numerous troubles. Problems to which actually have yet found a solution for. When I add myself via WordPress’s Friends/AcivityPub from my Friendica instance, I get errors. When I do the reverse from Friendica to WordPress, I get network failure issues, not network as in TCP/IP, but ‘Fediverse’ network issues, like it’s having issues with the protocol stack of this all.
So, explaining the details of what I have here. This whole server runs at my home server cluster with Proxmox VE managing the infrastructure as Virtual Machines, 3 physical servers, about to be adding 2 more once I rip them out of Kubernetes, will be added as additional nodes just to have higher availability. Within these 3 physical machines are 3 webserver instances, web1, web2, and web3. In pair to those I have redis in redis1, redis2, and redis3, these are running in sentinel master/slave replication, but to back even those, I have HAProxy handling the connection and routing to redis (master node) and Web1-3, and also my tunnel to CloudFlare which allows me to run this all nicely and securely with a tunnel I can just stop if the need arises.
Now this is a pretty complex setup, most would think, and in honesty, it’s not, not in the way I understand things at least. This is just a basic multi-node High Availability setup using a proxy server load balancer to handle things.
Where things seem to get complex is when it comes to the webserver itself. I run Nginx as my webserver primarily, but it seems this comes with lack of support for things like WordPress and Friendica which expect to make (bad) use of .htaccess files intended for Apache HTTPD. While it’s not entirely bad, the very specific intended use of .htaccess slows things down from having actual good configuration within the webserver itself. And, of course, only Apache HTTPD even reads that file.
So in further research, I need my own handler for /.well-known/webfinger, /.well-known/nodeinfo, /.well-known/x-nodeinfo2 and /.well-known/host-meta ? Hmm, okay, so I tried that with a line very much like this in my WordPress nginx config:
location ~ ^/.well-known/(webfinger|nodeinfo|x-nodeinfo2|host-meta) {
try_files $uri $uri/ /index.php?$args;
}
This, at least, looks right, and I got that from a resource that seemed to have an understanding of this, and actual known problems with ActivityPub and Nginx. But, once again, it seems to not actually be working, thus far.
At this time, I’m still working on solutions, and so I decided to open up the floor here with the issue, and see what I can get. Any comments? Any solutions? If you got any, leave a comment below!
– Psi-Jack
]]>To start, I investigated into the manual way to get into Kubernetes. See my one problem I always had with Kubernetes was storage. How does one handle storage in a clustered system, and actually get stuff done? I’ve tried a number of solutions for this over the years, and I’ve even implemented a few new approaches to some of them in this very project alone!
First time I really got into it, Ceph was badly supported in Kubernetes. Now it seems, it’s much more well supported, however, I still have some of the age old problems I had with Ceph, well on the side factor I am still running everything under Proxmox VE. Either way, I tried it with rook-ceph, even really did a number to get the external cluster well done, before I ultimately pulled it all out and ended up going with longhorn. Another can of worms there.
Ceph can be very nice, it was, in fact, for so very many good long years I used it. It’s around the time BlueStore came out it started becoming more beastly than my systems could really handle, than any systems could that wasn’t dedicated solely for the use of Ceph as a storage cluster. Don’t get me wrong, Ceph is amazing, Ceph is great. It just is a lot more demanding and not very forgiving if you don’t have what it needs.
A very interesting approach to storage. iSCSI under the hood for the equivalent to RBD from Ceph, bridging block devices directly to a system for a disk mount. That, at least, is for ReadWriteOnce type. If you go with ReadWriteMany as an alternative, is actually wraps that around NFS. Versatile, for sure, curious, and very noisy when it comes to kernel logs about disk I/O errors and such. I’m still using Longhorn today, but it’s definitely not without it’s own issues as well.
A very unusual choice, I know, but I’d long since switched my Proxmox VE systems from Ceph to GlusterFS, and it’s been a very surprisingly reliable choice for what I’ve been using it for. I do say, it’s been quite the interesting ride, though, trying to use this in Kubernetes. Fully backed out any support, and there’s 3rd party CSI drivers for it, also in the same boat, not exactly looking promising at all.
I did find one way of things though with it. Using NFS Ganesha for accessing GlusterFS, I was able to come up with some interesting ways around using GlusterFS with Kubernetes. Not the most ideal, mind you, but functional. I’ll be exploring more about this over time, it’s not without it’s own issues in itself. For lots of small files, as was needed for a Nextcloud kubernetes setup, it was so slow that it could never actually deploy without killing itself.
After I got my feet wet enough, I chose to start working with a template concept around Kubernetes, GitOps, Flux, and k3s. Something from the Kubernetes at Home (Now named Home Operations) Discord community, it’s rather quite impressive. I use it still to this day, and working with it, this has been one of the best setups I’ve seen using Kubernetes. Sure, I hit some bumps in the road here and there, but the overall setup is quite reliable and, for the most part, I could mostly nuke my whole cluster and rebuild it back up to what it is now, websites, applications, backups restored, everything. So, with this, I’d definitely have to go with, purely amazing concept!
You can look into this cluster here: https://github.com/onedr0p/flux-cluster-template
More to come on this adventure, I’m sure! Until then…
– Psi-Jack
]]>So I’ve been running the distro for several months now actually. I run it on my desktop, on my tabtop, and … No really that’s it. My other actually used computer is a MacBook Pro, so no Linux on there. So far, I actually am liking it! It’s stable, more so than I imagined. I imagined Arch is unstable, bad, painful, tedious….. Well, Arch Linux has more maintainers, has more actual testing, and more stability than I’ve seen in a while. Sure, some bumps may happen, but the problems are less common than I have experienced in the years past.
Of course I have all the trimmings, I have snapshots fully setup, I made a installation setup script for my install so I could re-use and setup my setup very much like I like, snapshots included.
Yep! That’s right! I’ve stepped away from Fedora, completely. Fedora is a good distro, but not really so much in the community sector very much. I literally waited months, heck more than that, years, to even try to get to be part of the Fedora team in some capacity. I decided enough is enough. They started removing packages that I used, just because they don’t maintain them anymore, like devilspie2, gone. I had it put back in by COPR, but I reported the issue, and literally nothing came of it.
So I’ve fully walked away from Fedora. I can do more with Arch in the community than any other distro, even not being directly part of the team. They actually accept community involvement more.
So, let’s see where this road takes us…
– Psi-Jack
]]>That is the question.
So to explain and translate these two as one of them is one I in particular made up based on what’s actually going on. VPN has been around for ages, and is generally used to bridge two private networks together, or at the very least bridge one private device into a private network, and do so securely with encryption.
Public services like NordVPN, ExpressVPN, ProtonVPN, etc. These are not actually bridging two private networks together nor are they bridging a private device to a private network. Instead what’s happening is you’re setting up an encrypted tunnel from wherever you are, be it home, a hotel or some random public WiFi, to an endpoint that then just funnels you out to the public internet again. So What is PVN? Public Virtual Network. It’s not really private like people think, because it’s consolidating a lot of people to one server to access the internet, much like your ISP does already.
A PVN can be useful in situations where you are on less secured private networks, such as a hotel or wifi hotspot, just to prevent local snooping of your internet access, but otherwise, a PVN is pretty much a general waste of money for more than needed.
To use it as an everyday purpose just to hide your personal traffic, from… Who? Your ISP? Your roommate? That’s not really worth paying $10/mo for or almost $100 up front for 2 years worth of limited speed access, just for this? Not to me it’s not. Not to mention, on that particular PVN you might also be dealing with a number of bad actors as well. Afterall, you’re all being funneled through the same outlet.
In short? Only when you absolutely absolu… No, simply put never, and you’ll learn why later on.
So, I don’t share a public network, but I do actually also use a VPN for it’s intended purposes, with the added benefit of also using it for other reasons I shall describe later.
I’ve primarily been running at any given point in time, an OpenVPN network for bridging my home directly to 2 or 3 VPS hosting providers, such as AWS, DigitalOcean, Vultr, and now Linode. This allows me to tie in my own personal servers from my home through their own private IP addresses of my server instances within these providers, safe and securely. This also helps in the simple fact that I don’t necessarily have to leave OpenSSH open to the world as well, because so long as I have a VPN connection open, I can use that tunnel to ssh in accordingly.
As for OpenVPN itself, I’ve used it quite a long time, it’s been familiar to me for so long. It has username/password support, it can do 2FA with some additional work, and so on. I even upgraded, finally, to Eliptic Curve profiles so I wasn’t using the heavy burdening RSA-2048 or even bulkier RSA-4096, and so throughput was definitely much faster.
I made a new change just recently though that had initially proven difficult, mostly due to lack of knowledge of this new thing, but also because it required more manual setup than I was used to having to do. That change was trying out Wireguard, the new simpler to set VPN solution everyone seems to talk about. And let me say this. Once setup and configured and templated properly, this solution is actually quite good these days. With NetworkManager supporting it, now also in GUI, and macOS, iOS, iPadOS, and yes also Windows (which I don’t use, hence last listed) all supported, and their clients are actually not bad either, this makes this leap a little interesting.
So, how much does it cost to run your own VPN? Well, that depends, but in general respects, for something simple and primarily personal use, you can either get away with doing this on a dedicated instance within a VPS you already have if you have one, for roughly about US$5.00/mo, with services like Linode, Vultr, DigitalOcean, that’s the recommended way to do it, but you could also potentially add it on to your webserver instance if you have one, and likely if you have a VPS you do. This is not as recommended, but it’ll do in a pinch if needed.
If you use this VPN heavily for literally every bit of your personal traffic, you might run into overages due to bandwidth… Just to keep that in mind.
So, I’ve been using OpenVPN for quite literally decades now, personally setup every time. From RSA 2048 to even RSA 4096, to more recently EC 384 which was an improvement to speed for sure. One factor of what I liked about OpenVPN was it’s versatility and ability to push IP’s, DNS, and Routes to the client all on its own. These are features you just don’t find in many VPN solutions.
I’m currently also running as of late a secondary VPN using WireGuard. This is a bit more manual, but the setup, once understood, is very simple. It’s just a step of setting up a key pair per client and IP per client, and such. It can actually be faster than OpenVPN, but I have not yet fully tested this, or other factors such as how well it works on mobile when traveling. So, details of all these will be handled in another blog entry.
For now though, that’s my story, and hope that the difference in VPN vs PVN (which is being mistakenly marketed as VPN), makes a difference in people’s understanding.
– Psi-Jack
]]>