Everything is changing. Faster and faster progression.

The new site does not look pretty at the moment but I want to get the domain pointing to the new site make development easier.

I currently have a front page, it is nothing fancy and simply has a redirect to the current blog after moving it to the hosting solution. This means that not all links will work correctly at this time. I am probably going to create a bug tracker simply to have the capability for my users to tell me things that they notice wrong with the site so I can slowly fix it all.

Surely some of the links used will point to the old edgelinux.wordpress.com that the site originally started out with, these are the sort of bugs that I will have to deal with. I have decided to not worry about the look of the site at this time and get the back-end services working nicely together first and then muck with the looks. I think that it does not matter how nice the site looks if the back-end sucks so that is the order I will work on the site.

I also got my Taxes done and that progressed my ability to buy the starter hardware 1 Odroid XU and 1 Odroid U3 for the build cluster I want to create. The time frame in which I am able to buy the hardware was pushed forward from 3-6 months depending on how much I would have saved per pay period. In the future the entire site may be hosted on the Odroid cluster on a business class internet line instead of going through a third-party host.

The reason for this is that there always seems to be something in every host that limits what you can do with your website. To cover what I would like to cover my needs would probably be a dedicated server which as per the monthly cost would start nearing or even surpass what it would cost for unlimited business class internet. The cons of this are that the speeds I will be able to afford are probably slower than If I went with a hosted package. Speeds of 25/5 are not the quickest in the world by any stretch of the imagination but I can always look for mirror services like SF to help on the bandwidth side of things were needed to keep the site loading quickly.

Since I am planning on moving the domain pointer to the new site, expect something to break temporarily. Expect entry portal to look basic and horrible at the present time. It is something that I threw together in a few minutes to point to the blog so I did not completely ruin current functionality of the website.

I also have other things that are in progress of changing at this moment that should allow me to do more on the tech side, I do not see any possibility resulting in me doing less considering I will have more time for tech. I will elaborate on what is happening in a later post, I have around a month and a half before this change actually takes place. Due to the time frame I do not feel I have enough accurate information to give you at this time, I am not sure exactly how things are going to progress. I just know I will not stop trying to improve my future which means the tech side will not be sacrificed.

A small update.

I’ve been working on the sub domains for Cutting Edge Linux the last few days. I am in a mix of the planning, testing and implementation stage of things. I have 3-4 services at the very least installed and tested whether their basic functionality is operational. I have not delved in deeper to do any fine tuning.

I have not worked on the look and feel of each site at all. They are basically just slugs waiting to be replaced or modified once the rest of the infrastructure is up correctly. Right now I have a copy of this blog, the starts of a wiki, a paste service, a url shortener and a rss feed aggregation tool.

Not too bad for just a few hours of work earlier today, I think this site is going to come together nicely in the end.

Site status.

To be honest I was not sure what was going to happen when I switched the hosting over to the new site. I purchased the hosting but did not transfer the domain over. I find this to be the best outcome at this time because then this can be my temporary site and I can work on redesigning a new framework and layout.

I can easily point to both sites at the same time by pointing to the new site in a crouton chroot that I use for a web development environment when using my Chromebook.

Migrating cuttingedgelinux.com

I am switching to a hosted solution vs wordpress.com. This is the first time I have used a paid hosting solution so there may be interruptions in cuttingedgelinux.com due to this. I plan on moving the site as it is and then doing an entire redesign.

cuttingedgelinux.com should redirect back into the old edgelinux.wordpress.com as a temporary measure. This does mean the CSS of the site may break temporarily until I get it migrated over. I am planning on temporarily doing a wordpress.org site since I already have the site on that information.

I want to refactor and redesign the entire site using another service completely that will allow me to provide a wiki, bug tracker and community portals of a variety of options depending on what I find required.

Cutting Edge Linux Manjaro Remix will not build on the KVM.

I keep maxing the capabilities of these VPS’s. So the KVM VPS has enough space to work with a net iso. The Cutting Edge Linux Manjaro Remix fails to build due to space limitations. This is unfortunate. It means a Cutting Edge Linux Manjaro Remix in on hiatus until I get a Local setup going of Funtoo.

I am planning on doing a local of Funtoo Linux install both on bare metal. Once I have a baremetal version of Funtoo I want to switch my virtualization environment to KVM instead of Virtualbox so I can use emulation techniques for ARM package testing.

The current plan.

Well I currently have all the stuff for the Manjaro build server developed on the KVM based VPS. I can login via SSH and run a single script to do the entire build of Manjaro. It operates very cleanly and works quite well.

I have been tracking the Parallella project for a while now, I thought I could do something extremely interesting with them so I asked some friends for their opinions. One of them said that the Parallella was essentially useless for what I wanted. Doing compiles over the Epiphany cores is not possible due to the way that they operate. Since the primary purpose is a cluster of build nodes for source based packages then this is not at all desirable.

Build server info & more.

Okay so my host kindly upgraded my inode count by 10% which enabled me to use 275k inodes. I still am not sure whether this is going to be enough to create an ISO. I was also advised that if I increase my space on the VPS that the inodes should increase accordingly and if not to ask them again for more inodes.

The error message that occured when I would run a pacman -S from within the chroot has changed. It now complains with essentially the same error except it was now / instead of /var/cache/pacman/pkg that was being complained about. After testing a reinstall of the chroot I am hitting the error with /var/cache/pacman/pkg again. The error first complained that the directory did not exist which it did, then it complained that I was out of space. After running dh -h; dh -i; dh -m everything looks good.

I’ve contacted my host again to see if I can get this resolved. The troubleshooting steps I have planned are to test this task both locally and on the KVM VPS. If it works locally and not on the KVM VPS I can safely assume it is due to the limit on HDD space in correlation to the inodes. At that point I can puchase some HDD space to determine if that is the issue or not.

build server inode issues!

I am having inode issues with the VPS that I am setting up for the build, I contacted my host to see if I can get the issue resolved. The issue is that I have maxed out my inode count when creating the chroot so it does not finalize correctly and Leaves no inodes for building the images.

I currently have used around 220k inodes and the cap is 250k inodes, it takes around 40k just to create the chroot due to the requirements of the filesystem.

I talked with the chat support and they advised me that the issue could be resolved but since I did not know how many inodes I would need the chat support agent advised me to email support about the issue and see what could be done on that end.

So that is probably why I was getting that /var/cache/pacman/pkg error and saying I was out of space. It was not space that was the issue but inode count.

I may have to purchase additional space in order to get more inodes, I am not sure what the resolution is yet. I will post back when I get word back from my host.

Manjaro build server progress!

Yesterday on my day off I went and had breakfast and coffee. I brought my chromebook with me to do some work or at least some research on getting a working Manjaro build server that ran on another Linux distribution. At first I was a little confused at what was required.

Originally I was advised that all I needed was pacman. Great I thought, pacman is in my repositories, I’ll get pacman installed and this will be done in no time!

I proceed to take a look into manjaroiso thinking I am going to be building this in the native environment, It became apparent quite quickly that it will be a large task in order to get everything installed that I required to do the build. Things were certainly looking bleak, It certainly seemed like I was venturing forth into dependency hell at this time. I was told that if the system has pacman it will work so I kept my head high and continued brainstorming.

Well if I cannot get this stuff going via the host system maybe a chroot is more appropriate! Trying to create a chroot I ended up destroying the host system. Chroot scripts I usually use/create clean up after themselves, This one did not and I rm -rf $CHROOT_DIR without even thinking about it. That was enough with /proc /sys and /dev mounted to kill the install enough that SSH would not allow me in anymore…

I reinstalled the Host and got it up and running yesterday/tonight. Tested the Chroot again and everything ‘almost’ worked. Everything seemed to install but I got errors when trying to install things with pacman saying it did not know where /var/cache/pacman/pkg was located. Usually this means the directory does not exist, I investigated with ls /var/cache/pacman/ and the pkg directory was shown. I also ended up with an error stating that the disk was full directly after the first error. The df -h command said the disk had 23 Gigabytes free.

I scratched my head, Founding this behavior sort of perplexing. I knew something must be missing but I was not sure what. I am still not sure what I am missing but I believe if I use pacstrap to set up the chroot that it may work correctly. The pacstrap command is not within the pacman package. I had to do my research into what Arch package contained pacstrap, Finally I found that it was the arch_install_scripts package that contained pacstrap. This was one of the packages that I had assumed I may have to build while getting overwhelmed by the dependency hell scenario stated above.

Everything required by this package was in the repositories so I decided to toss them onto my system. It appears that rather than having to build 10ish ebuilds that I would only require 1 in order to have pacstrap bootstrap the build directory correctly.

I still need to learn how to create a ebuild for this package, The dependencies are building as I type this and I am gearing up for learning how to create an ebuild!

If I am correct then this process is about 100x easier than I had originally calculated.

BONUS!

manjaro crouton renamed celce

Since I need something that is current in order to fix my crouton build (which I am considering doing soon.) I need to have access to a working install with merge tools and such.

Thus I have decided to rename my crouton fork celce. For lack of a better name it stands for Cutting Edge Linux Crouton Edition. I need to get some better platforms working though. ChromeOS does what I need at this time as a thin client but I miss having access to the Linux stuff!

XFCE + KWIN howto.

This is going to be a general how to that does not go into extreme detail. I am writing this because I mentioned running XFCE + KWIN in a thread a friend posted about the KDE framework being upgraded to use QT5. He showed interest in this setup and desired to know a bit more about it.

I will be explaining which packages are needed and some of the challenges of running cross-de software like this. Particularily because KWIN currently comes with some redundant software that can conflict when used with XFCE. The specifics of packages are based on Manjaro, customize them to what works with your chosen distribution.

In Manjaro there are 2 some what applicable packages. A KWIN stand-alone and kdebase-workspace which comes with KWIN and a bunch of extra stuff. If you think the stand-alone version is what you want then you are wrong, normally I am all for less bloat if it does not interfere with functionality. Which sadly using the more appropriate package does, The KWIN stand-alone module has no configuration tools at all. So you need kdebase-workspace which comes with the bloat of the activities stuff.

Now installing KWIN is only part of the decision in running KWIN within XFCE. You need to decide what to do with xfwm4. Do you keep it or remove it? I choose to remove it because it meant less bugs within my .xinitrc file used to launch the system through my DM.

Choice of DM’s. Well this is one of the most important parts of running a cross-de system. You want a DM that is easy to configure for custom use, The flashy fancy dynamic stuff may be nice when you run a ‘pure’ DE but when you are crossing the technologies it is a little different. I believe the other DM’s use .desktop files in order to perform their launchers. If you know how to modify these DM’s to work properly then go ahead. I opted for something a little more hacker friendly, Slim.

There are a variety of ways to launch KWIN which will yield differing results. You can load it in the DE itself and have it work its session magic. This is by far the most buggy method of launching KWIN. Why? Lack of control, You do not get to choose when things are run so they could be run in mismatched order and often times are. If you use sessioning accomplish this task then you will invariably end up booting into the system to have the sessioning kick in, retardedly load all your programs before your window manager and end up with all windows having their top borders under the xfce panel requiring you to use the ALT+LMB in order to move the windows. This is not the only reason I dislike sessioning, It directly causes many bugs when working with a cross-de system, eg. Every time I rebooted I ended up with Sessioning conflicting with a load-on-boot directive causing me to have an extra cairo-dock at every reboot. This bug was happening on my mothers box and I would go to use her computer, it would be slow as a dog and I would wonder why until I went to launch something in cairo-dock and she would have 7-8 sessions of it loading… Be aware that there is a bug in the XFCE sessioning set, the only way to eliminate this was to rm -rf the xfce sessioning directory before loading the environment.

The method that worked best for me was using slim + .xinitrc and loading kwin –replace and then xfce, this way ‘if’ the sessioning stuff does not bother you it will load your sessioned windows properly, not bugged as described above which happens because everything else is loaded before kwin is… If you choose not to remove xfwm4 you would need to load xfce and then kwin –replace.

There are also menu considerations when you mix technologies like this, Do you need to old WM settings, Theme/pointer stuff, etc, etc. Since kdebase-workspace is essentially a minimal KDE install this requires some modifications if you do not want redundancy that makes little to no sense. Two places to set start up programs… Which one do I use?

It is also possible to add the kde settings into the xfce settings manager.

Well I am sort of getting a little tired of explaining this at this time so I will leave you with an example profile I have created.

Wrong jumper!

The jumper I got is way too big for what I needed to coreboot my chromebook, I am going to have to keep an eye out for a jumper that will work!

Thus being the case when I have time I may work on manjaro crouton still. I am not happy about this though.

Funtoo, Manjaro crouton, Chromebook.

Funtoo has access to pacman without requiring unmasking of software so can do the same build stuff with manjaroiso easily. I believe the end result will be a more stable system for a server. I switched both the build server and my other server over to Funtoo in hopes that I can eventually learn how to use distcc, which distributes the compilation over multiple boxes.

ChromeOS will not operate the type of VPN that I want to use and with the breakages in crouton I do not know if it is worth maintaining long term. So I guess I am announcing that I am no longer going to maintain it, at all. If someone wants to fork it and merge the fixes that have been put in upstream it will probably fix it. So many things do not operate properly and it all seems like a daunting task where it started out as a fun task to see ‘if’ I could do it.

I have been seriously considering doing a coreboot on my Chromebook and just punting ChromeOS with the possibility of dual booting Chromium OS in the future if possible. This makes sense of why I am dropping maintenance for manjaro crouton, I will not be able to test it anymore as I did in the past. With newer chromebooks moving over to seabios the crouton project no longer makes as much sense to me as it once did.

I went to a computer hardware store, asked if they had a spare jumper I could have. I have all the equipment required to coreboot+seabios the chromebook and install some Linux distributions. I have done the research on how to perform the task both on the developers website and a few useful videos by johnny phung on youtube which I will link below.

It is recommended to fully read and comprehend all information available before attempting this task on your Chromebook. There is not a lot of information available and johnny phung is nearly the only person I could find with much information on this, his videos and posts lead me to the developers website.

Although the first link I will provide is to his Windows video it is relevent to the Linux side of things because he does not redundantly explain how to flash the bios with coreboot and seabios and he expands on a few commands that verify that everything is set correctly for the bios flash.

Coreboot windows video – Watch until 12:50. He uses alumium foil, I would only use a jumper for this. Got an old computer or hard drive sitting around? Grab a jumper off that!

Coreboot Linux video – There is the link for installing ubuntu but really any distribution would work, I am probably going to have 2-3 different installed distributions going or maybe 1 binary based and 1 source based. I have not decided yet what the plan is but I have a pretty good idea how to get a well compiled source based distribution to run on binaries as well and that would be quite ideal for a little Chromebook without much power to compile for itself.

 

Expansion on Arch OpenVZ VPS plan.

I want to expand on my vision of what the OpenVZ server will do, I am pretty sure it is possible to get manjaroiso working on something with access to pacman.

The reason I am setting up this server is upon request, nobody asked me to set up a server but people did want me to update my Manjaro respin, This video caused a whole bunch of people to request an updated spin. I do not see the demand for this changing in the future.

Any of you who know me know that I love tech but do not love when it eats up all my time, I like to expedite things so they go faster. Automating tasks is a great way to do this.

Gentoo VPS – installed, waiting on reboot.

I have completed the handbook a second time for the Gentoo install verifying that everything should be correct.  I am going to wait on a reboot due to the possibility of getting locked out of the server while at work.

At least I can use it from the ISO via a chroot and work on getting required software compiling!

I’ll post more info when I have some!