Hats off to Chapeau Linux (see what I did there?!) On the one hand it’s “just” a respin of Fedora -and an old version of Fedora at that (24, at the time of writing). On the other, not only does it have a pun in its name (chapeau=hat; fedora=type of hat; red hat=parent distro to them both, sort of), but it looks marvellous and has some incredibly sensible options for a desktop distro.
No Wayland is a big one that makes life easier. An attractive default icon set is another. Good multimedia codecs yet another. On the downside, however, it includes Flash, but you can’t have everything! It is also not using the latest and greatest Fedora as its base, true enough: but that probably makes for a more stable, robust experience, so it’s likely a ‘good thing’ overall. To take just one example of that sort of thing: it’s still using a 4.8 kernel as the latest kernel version available to it, which means you can install ZFS on Chapeau 24 in a way you can’t on ‘true’ Fedora 25.
Of course, as a former Englishman, Chapeau’s French connections are a bit of a worry (its codename is “Cancellara”, for example… who is a Tour de France stage winner; the ‘chapeau’ name is not entirely kosher English, either!) But perhaps we should think of it as delicious exoticism instead of anything to carp about! On the other hand, I would pay good money for a distro with a codename of Agincourt! (Just joking! Crecy will do).
As a custom respin, it is to be expected that its user-base will likely be measured in the tens or hundreds rather than the thousands. At the time of writing, for example, it’s not even listed on the DistroWatch popularity charts. But I find that inexplicable, really: if I was contemplating making Fedora my main desktop distro, I think I’d at least give Chapeau a look first. I definitely found it a better out-of-the-box experience than Korora, another Fedora respin, for example: for one thing, it installed in VirtualBox just fine!
Anyway: given it is fundamentally Fedora-plus-tweaks, it should be no surprise that Atlas and Oracle work on Chapeau 24 and up perfectly well.
As is the case with all Atlas/Oracle installs, your server needs to be built with at least 5GB RAM and at least a 40GB hard disk; you must additionally ensure you have installed wmctrl or xdotool before attempting to launch Atlas.
Please watch the generic Atlas videos here, here and here to get an idea of how Atlas works in general. The generic documentation (including those videos) are available from this page. Below are Chapeau-specific notes.
2.0 What’s been tested?
- Chapeau 24
3.0 Operating System media
The Chapeau project’s download page is my source for the installable DVD ISO. There are no varietal flavours to worry about: there’s just a single Gnome live DVD ISO to get.
4.0 Operating System installation issues
Chapeau installs without drama and you can pretty much take all the default suggestions without alteration.
You start by booting into a workable ‘live’ environment direct off the DVD, from which you take the ‘install to Hard Drive’ option:
Clicking that install option then launches something that looks almost identical to the standard Fedora installer. As with that standard installer, most options are set for you automatically and you can just accept them. You are required to confirm where you want the O/S installed to:
So that requires clicking on the ‘Installation Destination’ item -but you can immediately click the ‘Done’ button after that without having to do anything other than accept the default partitioning options. Atlas will find them fine.
If you like, you can click on the Network and Host Name option, too (though the installer doesn’t require you to do so). If you do, you get to specify a sensible hostname for your new server. If you choose not to, your server acquires the singularly unimaginative “localhost.localdomain” name, which is something Atlas will then sort out for you later on.
Other than that, I was, for the most part, able to accept all the defaults offered during the installation. However, I did make one small change to the default offering here:
Specifically, when creating the non-root user for the system, make sure to check the Make this user administrator option. It means you will be able to sudo -i to become root (requiring only that you know your own password) instead of having to su – root (which requires that you know root’s password). It’s a minor thing, but I prefer it that way.
Otherwise, the O/S installation is straightforward. Once the server comes back from its first post-install reboot, you may need to bump the screen resolution up. In VirtualBox, for example, you can only get to 1024 x 768:
…and that’s not enough for Atlas. For VirtualBox, at this point, I had to install the Guest Additions. That requires a number of steps.
First, as root (or prefixing with ‘sudo’):
dnf -y update kernel* reboot
Wait for the PC to come back up, and then as root (or sudo) once more issue these commands:
dnf -y install kernel-devel dkms
Once that lot is installed, click to install the guest additions and either click on the option to auto-run the Additions installer, or (again as root) use these commands to run the installer manually:
sh /run/media/<username>/<VBOXADDITIONS_version>/VBoxLinuxAdditions.run reboot
However, doing things this way causes a lot of new software to be installed, such that you are no longer working on a ‘clean’ distro at the point that you run Atlas. It nevertheless works OK, but if you could avoid installing stuff onto your clean distro, it would be better to do so. Chapeau installations using VMware, for example, do not need this ‘install lots of stuff’ approach before they can use a high screen resolution.
Once the Guest Additions are installed, all sorts of screen resolutions then become available to VirtualBox users:
Once you get the screen resolution sorted, you can install wmctrl or xdotool without fuss:
dnf install wmctrl
5.0 Running Atlas
Assuming you’ve got a good screen resolution (better than 1200 x 750) and have installed wmctrl (or xdotool), you can download and run Atlas in the normal way:
wget http://bit.do/dizatlas -O atlas.sh chmod +x atlas.sh ./atlas.sh
At that point, you just start following the prompts. Because this is (fundamentally) Fedora, there will be a very lengthy delay in working out what distro you’re working on (Fedora being one of the few distros that don’t include the lsb_release package by default). But be patient, and you’ll eventually be told:
If you didn’t take the opportunity during the O/S install to alter your hostname, Atlas will now spot that and demand you do something about it:
If you just press [Enter] here, you’ll get given the default suggestion of ‘osrvr.test.lab’. If you want a better name than that, just type a fully-qualified hostname (i.e., with a domain name component separated by periods/fullstops). If you don’t have a proper domain name, just make one up: ‘localdomain’ will do just fine, for example. As you can see here, I tend to use ‘something.dizwell.home’ -and “dizwell.home” is not valid outside the walls of my study, but works fine.
Other than this, just follow the usual Atlas prompts until it completes and triggers a reboot of your new Chapeau server.
6.0 Installing Oracle 12c
As with many ’boutique’ distros, when you start the Oracle installation, you’ll be told your server doesn’t meet requirements:
It is fine to click [Yes] here to ignore the warning, though, and proceed with the installation regardless.
Otherwise, you can mostly click [Next] through the installation wizard and nothing very alarming will happen at all! There are, for example, no compilation errors during the installer’s linking phase.
There are no compilation errors raised during the installation’s phase.
As a result of there being no linking errors, an atlas-postinstall.sh script is created in the oracle user’s Documents directory. This can be run once the linking phase is complete (or after the entire installation, including database creation, completes), if you would like to do so. Running it is entirely optional.
Since I am my own oracle user, you here see me travelling to my own Documents directory and discovering the existence of a script called atlas-postinstall.sh. Note that it is already executable: Atlas created it that way. So all I have to do is type:
…to run it. Nothing appears to happen when you do, and there’s no indication of success or failure: but it has done its job anyway. Its effects can be seen in this screenshot:
The return from that simple SQL query is displaying quite nicely in the terminal session entirely because the post-install script has set decent defaults for page- and line-size in SQL*Plus. Without it, the results would be visually scrambled to the point of illegibility.