dutch-master

You may want to start afresh with emerge --jobs=4 --load-average=2 --emptytree --with-bdeps=y -auDN @world Worked for me last time I got stuck. ? Note: takes a long, long time on old(er) hardware! You may want to remove some large packages like the DE (Mate, Gnome, KDE) as well as mesa, xorg and similar to speed up the process.

Apols, I hadn't checked, but I got the same results as you got: 404 ? Alternatively, use the generic-64 stage3 file (it's there, I checked ? ) then change the cflags and chost entries in make.conf after rebooting into the actual system: CHOST="x86_64-pc-linux-gnu" CFLAGS="-march=amdfam10 -O2 -pipe" CPU_FLAGS="3dnow 3dnowext mmx mmxext popcnt sse sse2 sse3 sse4a" Then do a world update: emerge --jobs=4 --load-average=2 -auDN @world It may not be the "official" way, but at least it's a start ?

Have a go from this site: https://build.funtoo.org/1.4-release-std/x86-64bit/amd64-k10/ HTH!

Managed to install lxd and everything that comes with it on my AMD k10 system (Phenom II X4 965 proc, 16GB RAM, 120GB SSD). Then I proceeded by getting the lxd image for k10 arch and set up lxd following the Wiki by DRobbins. but launching the testcontainer as per the instructions failed with an error. After some trial and continues errors I managed to get both the lxd-generic_64 and intel westmere images installed and associated with a testcontainer: # lxc list +---------------+---------+------+------+-----------+-----------+ | NAME | STATE | IPV4 | IPV6 | TYPE | SNAPSHOTS | +---------------+---------+------+------+-----------+-----------+ | testcon2 | STOPPED | | | CONTAINER | 0 | +---------------+---------+------+------+-----------+-----------+ | testcontainer | STOPPED | | | CONTAINER | 0 | +---------------+---------+------+------+-----------+-----------+ testcontainer is associated with the generic_64 image, testcon2 has the intel westmere image as its base. Attempts to launch these fail with an "error: not found" message. The logs for both are identical (except of course for the name) and show this: lxc info --show-log testcontainer Name: testcontainer Location: none Remote: unix:// Architecture: x86_64 Created: 2021/02/28 20:51 UTC Status: Stopped Type: container Profiles: default Log: The "location: none" instance is suspicious, so I had a look in /var/lib/lxd/containers: ls -l /var/lib/lxd/containers/ total 0 lrwxrwxrwx 1 root root 54 Feb 28 21:53 testcon2 -> /var/lib/lxd/storage-pools/default/containers/testcon2 lrwxrwxrwx 1 root root 59 Feb 28 21:52 testcontainer -> /var/lib/lxd/storage-pools/default/containers/testcontainer And a look inside /var/lib/lxd/storage-pools/default/containers/ resulted in: ls -l total 0 d--x------ 1 root root 78 Feb 28 21:53 testcon2 d--x------ 1 root root 78 Feb 28 21:52 testcontainer I see a distinct lack of permissions, is that the root (sic ? ) cause of the issue? Anything else I missed or went wrong? TIA! [edit: I forgot to mention that during lxd init I choose btrfs as storage backend, the relevant btrfs packages have been installed (including the maintenance package) /edit]

Oh, apols, didn't catch that ?

Sorry to see you stop JFS support, as I use it pretty much exclusively. (not just Funtoo, but on my Devuan installs as well) At the time I choose JFS over XFS as the former had slightly better performance for small file storage, while being on par for large files. What does this mean though for JFS support in Funtoo? Will JFS remain an option to install on? Or is this just for a personal project you run?

Did you sync your local repo with the online one? Do ego sync Then run the emerge command again.

Could a dedicated fork for -bin packages be an option? A bit like a "Debianized Funtoo", just -bin packages to install. Great way to get going on older hardware! Case in point: my CPU dates from 2012-ish and I'm not in a position to upgrade to a Ryzen system any time soon. Building libreoffice-bin takes up a lot less time then libreoffice, which is often newer then the -bin package, but libreoffice-bin doesn't always build successfully whereas libreoffice usually succeeds. Other big packages like mesa, ffmpeg etc also eat enormous amounts if install time to get a fairly basic working desktop. Having said that, installing a bare basic Funtoo system w/o any graphical environment (no X nor a DE) is pretty fast, usually in about 1 hr on my ancient hardware. (AMD A8-5600K APU, 32GB DDR3 RAM, GeForce 710 video card. Hey, I'm not a gamer, it works! ? ) Well, I s'pose it'll all come down to a effort vs impact analysis for you and the dev-team: how to get the best possible impact for the most amount of users in the least amount of time. I'd say you guys do a fine job on that ?

Ah, many thanks again. I'll retrace my steps and see if I can get a working system this time ? (crosses fingers!)

Well, that didn't work. Blocked the 1.8 release globally in make.conf but the 1.4 version it wanted instead is no longer available. For some reason, the libvpx line was missing in funtoo-cautionary (at least on my system) so it wasn't blocked to start with from that layer. No idea where the blockage comes from then. As I was quite fed up by now, I tried reverting the 1.4 release back to 1.3 by first removing everything outside the core flavour, setting ego.conf back to 1.3, then deleting the existing package cache (and rebuilding it with ego sync) before running emerge -auDN @world. To be greeted by a glibc error that refuses to downgrade my system. Regrettably, it appears Funtoo 1.3 is no longer available as I can't find any stage3 or ftp-server directory for it any more. Essentially, if you want to install Funtoo+Mate, you can't. Libvpx blocks it on 1.4 and 1.3 is no longer available for a workaround solution. Installing a (popular) DE shouldn't require a user to jump through so many hoops ?

Ah thanks, I wasn't aware there was a bug filed already (didn't think checking the bugs list). I'll have another go later.

Reviving my attempts to install Funtoo 1.4, I've run into an issue with libvpx-1.9.0. This package is masked by package.mask but it is required for installing Mate. Several packages depend on libvpx, including ffmpeg, gst-plugins-[libav,meta], brasero and mate (meta-package) itself. Is there an alternative/workaround/fix for this? TIA!

Belated (attempted) answer: You can use 64-bit Linux on 2GB main-board systems. I happen to have a Shuttle All-in-one system with an Atom D525 as well and 64-bit Funtoo compiled fine. Takes a looooonngggg time, but it works.

Welcome (a bit belated, but still ? ) Given that you currently have no working system, consider starting afresh with the Funtoo install-guide: https://www.funtoo.org/Install Don't try short-cuts, take it step-by-step. Keep your make.conf as clean as possible, likewise epro mix-ins. Alternatively, use emerge to remove all X-related packages from your current system. emerge -aG xorg-x11 xinit xdm emerge -a --depclean Let it finish, then reboot. Emerge uses a separate build-directory for each package, including log and build files. Any issues emerge runs into that fail building the package will be reported there. You'll need these when reporting the error/failure.

The Wiki page on Mate (<- link!) stipulates it needs dbus as a prerequisite, so: rc-update add dbus default rc HTH!

OK, performed the usermod command again, minus the plugdev group, tried again to log in remotely and now it works as it should. Thx for the pointers!

The output is only the group name of the regular user. I had performed the usermod command as per the wiki install page, but it bombed out with an error about plugdev group not existing. I'll redo the usermod command w/o the plugdev group tomorrow. In 4 hrs the alarm goes to get to work... ?

Right, this is pretty annoying, Just succesfully installed a fresh Funtoo 1.4 (on some not-so-fresh hardware: Core2 Duo, 4GB RAM, 120GB SSD) and rebooted. - bare metal, not a VM! - logging in locally on the cli works (as root) - remote login (ssh) as normal user works - elevating permissions to root (su) fails - I've had this on a different system (Devuan Beowulf) and the recommendation was to revert to the old Unix way: su - - using su - fails with "permission denied" error, repeatedly - alternative was to add ALWAYS_SET_PATH yes into /etc/default/su. I did so, no change in behaviour. Reboot, tried again, again, no change in behaviour, I still can't elevate permissions to root remotely. Unless this can be solved, I'd have to revert this system to Devuan Ascii, 'cause I know it worked on that. Suggestions? TIA!

Sorry to report, but you have a major problem there: the link you supply refers to a local file (see the 192.xxx.yyy.zzz IP address range) so it's not available to anyone except yourself. Not very useful for sharing ? You're also advised to translate the download page in English, so people know what it is they're looking at before committing a download. HTH!

Do ego sync emerge -auDN @world Evaluate the list: how many packages require a rebuild and how much of those are large programs/tools/libs (like mesa, ffmpeg, etc)? If there's a fair amount, it might be quicker to re-install from scratch. I noticed you have an core i7 proc, make sure you set the MAKEOPTS flag to a suitable number (17 should do it) MAKEOPTS="-j17" You can put that in make.conf as it's a global flag. HTH!

Excellent! Great news, I'll investigate further tomorrow. Many thanks helping out! :)

Thx! Yes, I've found that (contributed the installation part for it) but it doesn't really answer how to create images myself. Is it just a text file, if so, what syntax do I need and if not, how do I create an image, especially for Devuan while running Funtoo.

Did you find this page on the Brother website yet? http://support.brother.com/g/s/id/linux/en/index_rp.html?c=gb&lang=en&prod=mfcl2710dw_us_eu_as&redirect=on Go to "Driver Install" -> "Cupswrapper driver install" -> scroll down to step 5 a/b.

Before asking my list of questions, first some back ground info: Since earlier this year I work for a thrift shop-chain. Part of my job (for 2 days/week) is running the PC refurbishment project. When I first joined this project Linux Mint was used to install a new OS on refurbished computers. Trust me, systemd on a P4 is a disaster. But who wants systemd anyway ;) So after a while I came across some nice professional network hardware (Gbit managed switches, several) and using an Intel Core 2 machine I created a local network w/o internet access. I also took the decision to change the OS to Devuan Linux with the MATE DE, as the intended audience/user-base are novice Linux users. To this aim I maintain a local repo of Devuan, updated weekly via a portable HDD drive. In the near future a new fiber-optic internet connection will be activated, but as its speed cannot match what I have now to transfer large chunks of data (sorry, packages ;) ) I'll keep the local repo in use for the time being. But the updating process will need some overhauling, having to log around with heavy USB drives is getting quite tedious now. So, with the upcoming access to high-speed internet I've obtained permission to connect to the to-be-launched wifi network during the night to update said repo. But I also would like to have a local Funtoo repo on the same machine, equally updated during the night. This means I need at least 2 virtual machines and from reading my way around, it seems Linux Containers may do what I want: run a Funtoo base system, with a pair for virtual machines on top, one running Funtoo and the other Devuan, each updating itself during the night before switching off again. (I'll be writing some scripts for that, but first getting to grips with LXC!) Is this possible, how do I proceed (the replacement server already has a fresh Funtoo 1.2 install on it) and how do I create my own containers as I'm aware there's nothing for Devuan nor the Core2 metal the server runs on (Funtoo only supports Westmere so far, by the looks of it). TIA!

Try installing Cups and manage your printer from there. Also, mind that Brother has excellent instructions on how to connect a networked device, including the correct URI syntax on their website.