lazlo.vii
-
Posts
61 -
Joined
-
Last visited
-
Days Won
6
Content Type
Profiles
Forums
Blogs
Posts posted by lazlo.vii
-
-
I have version 3.6 installed.
-
I don't know what I did, but what ever it was it wasn't good. I had to shut it down my server and move it to clean the floor and after booting it back up I noticed my containers were not running.
Running /etc/init.d/lxd start:
* Starting lxd service ... * start-stop-daemon: /usr/libexec/lxd does not exist [ !! ] * ERROR: lxd failed to start
I tried "rc-update del lxd default" and rebooted then "rc-update add lxd default && openrc"
Still had the same issue.
So then I tried "mkdir /usr/libexec/lxd && openrc" and I got:
* Starting lxd service ... [ ok ] lxc list Error: Get http://unix.socket/1.0: dial unix /var/lib/lxd/unix.socket: connect: no such file or directory
tail -f /var/log/syslog showed me this:
/etc/init.d/lxd[21723]: ERROR: lxd failed to start /etc/init.d/lxd[21771]: start-stop-daemon: failed to exec `/usr/libexec/lxd': Permission denied
If I start lxd as root with "lxd --group lxd" I can get my containers but I have to do it in a tmux session so I can leave it running after logging out of ssh. Doing this I am unable to tell any difference from the way my server used to start lxd via openrc.
What ever I did wrong was done days or maybe weeks ago. Did I miss configure something? Did I use the wrong option in etc-update? I have tried recompiling lxd, lxc, lxcfs, and cgmanager but it didn't change anything. I am at a loss. What should I look at next?
-
I did this a while back to set up an AP with hostapd:
/etc/conf.d $ cat net.brwlan template="bridge" ipaddr="192.168.0.201/24" gateway="192.168.0.1" nameserver="192.168.0.1" slaves="net.eth0 net.wlan0" stp="on" forwarding=1
with
/etc/conf.d $ cat net.eth0 template="interface-noip"
and
/etc/conf.d $ cat net.wlan0 template="interface-noip"[/code]
Now, since you already have an AP I think you could configure net.brwlan to connect to the AP but to be honest I don't know much about it since that is not documented in the networking guide. NetworkManager seems to be the only official way to connect to an AP in Funtoo. If I had to do it I would start with reading up on wpa_supplicant and try to write a bash script to bring the bridge up and and then connect to the AP.
Or, I would try to let NetworkManager manage the wifi and then attempt to get my box to just act as a router with no bridging of NICs at all. I don't have a lot of faith in NW though. Unless it has improved greatly in the past 5 years I think it would just get in the way. Have a look at https://wiki.gentoo.org/wiki/Home_router
It might give you some ideas.
With luck someone that knows the Funtoo networking scripts more intimately than I do can give you a better answer.
-
If you can tell us what you are trying to do exactly it could help us give you the best answer. If you want to run a wireless access point there is software that can help with that, like hostapd. If you just want a simple bridge to play with you could take a look at https://www.funtoo.org/Networking just don't add any bridge you create to a run level with rc-update and it will only start when you start it.
As with anything in Linux there are going to be a lot of ways to do it but really the best way is going to depend on what you are trying to do and why.
-
7 hours ago, palica said:
this is kernel related IMHO.
are you using hardened stage?
what kernel version ...
I am using the stage3 from here:
https://build.funtoo.org/funtoo-current/arm-32bit/odroid-xu4/
The kernel is 4.14.78+ from here:
https://github.com/hardkernel/linux/tree/odroidxu4-4.14.y/kernel
I am using the default kernel config for an Odroid xu3 or xu4. The kernel was compiled following this guide for native compile:
https://wiki.odroid.com/odroid-xu4/software/building_kernel#native-build
EDIT: I copied the kernel .config to http://termbin.com/zeue
-
* Messages for package sys-devel/gcc-7.3.1: * Failed to set XATTR_PAX markings -re /var/tmp/portage/sys-devel/gcc-7.3.1/image//usr/libexec/gcc/armv7a-hardfloat-linux-gnueabi/7.3.1/cc1. * Failed to set XATTR_PAX markings -re /var/tmp/portage/sys-devel/gcc-7.3.1/image//usr/libexec/gcc/armv7a-hardfloat-linux-gnueabi/7.3.1/cc1plus.
This shows up at the end of merging gcc and there is a similar message after merging python. Is it going to impact my system stability or security in the long run? I would like to go ahead and start building my X server, DE, and other things I had planned for my XU4 but I don't want to waste time. With a bare bones stage3 install it took over 6 hours just to "emerge -e @world."
-
Today I decided to start fresh with the stage3 tar from late July. I have made good progress in that gcc 7.3.1 compiled without crashing and while respecting my MAKEOPTS settings. It only took it about 3 hours to run. Now I am working my way through the rest of the 1.2 upgrade and am currently running "emerge -e @world" but there is one thing that I don't like.
When an ebuild is done compiling I get an error that looks something like this:
"Failed to set XATTR_PAX markings <string for markings> </path/to/portage/build/dir>
Life is never dull as long as I have my Odroid ?
-
Thanks for the tips. Is there any significant difference between your suggested EMERGE_DEFAULT_OPTS and MAKEOPTS="-j2 -l4"? Right now my make.conf has MAKEOPTS="-j4 -l4" and the gcc ebuild was still spawning enough threads to send the load average through the roof. I tried everything else I could think of short of editing the files in Portage. I didn't try that because I know that I don't know enough to do it right and I would hate to screw up my system with an ugly hack 'n' slash of Portage. I even set up an NFS share to hold my root Odroid file system so I don't have to worry about corrupting the SD card with the constant I/O of compiling packages. Well, filesystems, really because it is hosting the default ubuntu as well as funtoo, devuan, and armbian. I am just one file edit and reboot away from switching between them and the gigabit NIC is faster than the SD card reader by a long shot.
Anyway I'll try the EMERGE_DEFAULT_OPTS later and see if it has the desired effect. Thanks again.
EDIT: OK, the EMERGE_DEFAULT_OPTS line you gave me is pretty much the same as "emerge --jobs=2 --load-average=4" and it had no effect on the amount of threads portage was spawning. I guess I'll either wait until an official update comes down the pipe or get brave enough to try editing Portage files.
-
Thanks for adding the 2FA to the site. I wish more would do it. I use it every where I can. Even my home server has the google-auth pam modules for ssh.
- cuantar, NikosAlexandris and drobbins
-
3
-
I masked =sys-devel/gcc-7.3.1-r3 and tried the upgrade, but has to revert to a backup. It seems that the gcc 7.3.1 is ignoring MAKEOPTS and spawing so many threads the my poor xu4 segfaults the compilation or hangs and needs to be reset. The freeze up has got to heat related since the stock fan is always at max rpm when it happens.
-
I have an Odroid-xu4 and I would like to install Funtoo. The newest stage3 file is from late July and I took a peek inside at the ego.conf file. It isn't setup for 1.2 and naturally the version of ego is less than 2.6.0.
Are there plans to continue supporting the Odroid-xu4? I understand that with limited resources you guys can only do so much. If the Odroid support is "on the back burner" that is ok. If it is discontinued I can live with that too. I'll just look for another solution. Either way I would be grateful for a status update.
-
@cardinal
>ego query v ego app-admin/ego| slot| repo --------------+-----+------------------- 2.4.2| 0| core-kit/1.2-prime 2.5.0.2| | core-kit/1.2-prime 2.5.0.4| | core-kit/1.2-prime * 2.6.0-r2| | core-kit/1.2-prime 9999| | core-kit/1.2-prime
>equery d grub * These packages depend on grub:
On my server Portage thinks that nothing depends on grub.
2 hours ago, Oleg Vinichenko said:This is link to the new Funtoo Installation Guide. "This page was last edited on October 5, 2018, at 22:02."
-
I followed the instructions in this thread:
Now after running "ego sync" and "emerge -auND @world" when I run "emerge -cp" I get the following output:
>sudo emerge -cp Password: * Always study the list of packages to be cleaned for any obvious * mistakes. Packages that are part of the world set will always * be kept. They can be manually added to this set with * `emerge --noreplace <atom>`. Packages that are listed in * package.provided (see portage(5)) will be removed by * depclean, even if they are part of the world set. * * As a safety measure, depclean will not remove any packages * unless *all* required dependencies have been resolved. As a * consequence of this, it often becomes necessary to run * `emerge --update --newuse --deep @world` prior to depclean. Calculating dependencies... done! >>> Calculating removal order... >>> These are the packages that would be unmerged: sys-boot/grub selected: 2.02-r1 protected: none omitted: none sys-boot/efibootmgr selected: 15 protected: none omitted: none media-libs/freetype selected: 2.8 protected: none omitted: none sys-libs/efivar selected: 35 protected: none omitted: none All selected packages: =sys-boot/efibootmgr-15 =sys-boot/grub-2.02-r1 =sys-libs/efivar-35 =media-libs/freetype-2.8 >>> 'Selected' packages are slated for removal. >>> 'Protected' and 'omitted' packages will not be removed. Packages installed: 434 Packages in world: 22 Packages in system: 81 Required packages: 430 Number to remove: 4
I have not removed any of the packages listed in the output and wanted to ask if this is expected and intended behavior. Pulling out grub seems a little bit reckless. If I remember correctly, following the Funtoo Install Guide caused grub to be pulled in as a dependency of boot-update. Perhaps the upgrade instructions for ego 2.6.0 should be expanded to cover this.
In the meantime I'll just add sys-boot/grub to my world file.
-
OK, I have do my tests and this is what I have found:
1.) While the script in the Gentoo wiki will generate pre-install snapshots it will not generate post-install snapshots.
2.) As I feared, the script in the Gentoo wiki is triggered per ebuild install phase. This means that you get one snapshot per package Portage installs. This is a bad thing because snapper limits the amount of snapshots it stores and in a large emerge run you might end up overwriting some of the snapshots you made just a while ago.
For this to be really effective Portage would have tell snapper to make a pre snapshot during the pre-compile phase after it does all of it's sanity checks. Then after the last package is installed and about the same time Portage gives the output messages from all of the packages it would need to call on snapper again to make it's post snapshot.
Snapper does do timeline snapshots correctly with cronie installed though, so that is about 75% of what I want right there. So from here I'll start backing up my server and get it reinstalled with a btrfs raid10. I am unsure of how to go about officially requesting the changes I desire or if these changes are really feasible. If anyone working with Portage cares to comment I would be grateful to hear your thoughts on the subject.
-
8 minutes ago, shadow53 said:
I found this page on the Gentoo Wiki, which outlines a way to take pre- and post-snapshots when using Portage. Haven't tried it yet.
Thank you very much for the link. Tomorrow I'll spin up a new Funtoo VM and test this out. If it works as expected is the proper procedure to file a bug against btrfs-progs for a use flag to pull in snapper on request and another bug against snapper to add a use flag for portage auto snapshots via /etc/portage/bashrc?
-
Hi,
I hope this is the right place to post this.
A bit of background that can be skipped if you are in a hurry:
I am by no means a programmer and try hard just to stay out of the way and keep my head down when using open source software. I've been using Funtoo for a long time now on my home servers. These systems started as lab style "blow them away at the drop of a hat when things go bad while flying by the seat of my pants learning a thing or two along the way" machines but over the years they have evolved into a much more stable pattern. I have a desktop for gaming but I don't run Funtoo on it, choosing Mint instead because "It just works" most of the time. One of the things that bothers me about Mint is the lack of real data protection in the default install so I set up a raid like I do with my server but instead of mdadm I used brtfs, This means that the default filesystem snapshot tool in Mint (TimeShift) can't work on my system. So I dug around and found a very cool utility called "snapper" that auto snapshots my btrfs filesystem. It does this transparently in the back ground and can be configured to keep so many hourly, daily, weekly (and so on) snapshots. The cool thing is that it makes a pre and post snapshot every time I unistall or uninstall software with apt. I looked for snapper in portage and it's there. This has me thinking about starting a fresh install of Funtoo on my server and using btrfs raid10 and having snapper as a second layer of system protection.
So to the point:
Can snapper automatically snapshot a btrfs filesystem before and after an emerge command is executed? If not would making it do so be trivial or non-trivial? If it does so or can easily be made to do so why not add a snapper use flag to btrfs-progs so it would be pulled in if desired by the system admin?
This is the kind of idea I normally keep to myself because I can't write it and I tell myself "If it was that good of an idea it would be done already." In this case it's something I really would like to have and it would benefit so many people using both Funtoo and Gentoo that I felt I should at least ask.
cannot set profile
in Desktop Help
Posted
I have to ask, are you running the command as root or at least using sudo as a non-root user? If not you can't change files in /etc