View previous topic :: View next topic |
Author |
Message |
Albion n00b
Joined: 07 Feb 2012 Posts: 31 Location: England
|
Posted: Tue Jun 03, 2014 5:56 pm Post subject: Chapter 6 errors |
|
|
Good hello, all,
I just got through this stage:
Quote: | # chroot /mnt/gentoo /bin/bash
# source /etc/profile
# export PS1="(chroot) $PS1" |
but when I typed in the 'emerge-webrsync' command the system responded by scrolling through a long list of files that said things like:
Quote: | Typing to receive 20140424 snapshot from http://distfiles.gentoo.org ...
Fetching file portage-20140424.tar.xz.md5sum ...
Fetching file portage-20140424.tar.bz2.md5sum ...
Fetching file portage-20140424.tar.gz.md5sum ...
20140424 snapshot was not found
|
The above list is the bottom of the list because there's a list longer than my screen will show that shows the same think, except for the fact that the dates are shown in descending order where the date shown above is at the end of the list.
ping -c 3 www.gentoo.org results in "ping: unknown host www.gentoo.org"
Before I chrooted into the new Gentoo environment I was able to access the net because I used wget to download the stage3 tar file. At this time, lspci -k reported the NIC module in use was alx (I have an MSI Gaming motherboard with the Killer NIC) but after chroot lspci is an unknown command. Running ifconfig shows eth0 has disappeared and has been replaced with "enp3s0". I don't know what's going on because I've gone through the manual before and did a temporary installation of Gentoo which completed successfully. I just started again to become more familiar with it and learn more.
Can anyone shed any light on this problem?
Thanks in advance.
A. |
|
Back to top |
|
|
cwr Veteran
Joined: 17 Dec 2005 Posts: 1969
|
Posted: Tue Jun 03, 2014 6:21 pm Post subject: |
|
|
enp3so is now considered a much better name than eth0. The easiest way to deal with it, as far as I am concerned, is to add "net.ifnames=0" (no quotes) to the kernel command line. Then you get the standard eth0 / wlan0 names back.
lspci isn't part of the Stage 3, AFAIK, so it won't exist until you emerge it. The first problem is to get the network up (though in the last resort you can use sneakernet to load /etc/portage/distfiles). I'd attack the network problem next, and ask for help here giving details of error messages from eg: ping.
Will |
|
Back to top |
|
|
freke l33t
Joined: 23 Jan 2003 Posts: 977 Location: Somewhere in Denmark
|
Posted: Tue Jun 03, 2014 6:22 pm Post subject: |
|
|
You did remember to copy over resolv.conf?
(can you ping IPs ie. 8.8.8.8?)
Code: | Code Listing 1.3: Copy over DNS information
(The "-L" option is needed to make sure we don't copy a symbolic link)
# cp -L /etc/resolv.conf /mnt/gentoo/etc/
|
|
|
Back to top |
|
|
Albion n00b
Joined: 07 Feb 2012 Posts: 31 Location: England
|
Posted: Tue Jun 03, 2014 7:27 pm Post subject: |
|
|
Thanks.
Ping 8.8.8.8 works
The previous time I went through the manual I wrote it down step by step. I've been working from both the manual and the quick install guide. I noticed a few problems with that so when I've got my installation up and running I'm going to update that. I want to give back to Gentoo
A. |
|
Back to top |
|
|
Hu Moderator
Joined: 06 Mar 2007 Posts: 21635
|
Posted: Wed Jun 04, 2014 1:52 am Post subject: Re: Chapter 6 errors |
|
|
Albion wrote: | Running ifconfig shows eth0 has disappeared and has been replaced with "enp3s0". I don't know what's going on because I've gone through the manual before and did a temporary installation of Gentoo which completed successfully. | Congratulations, you are a victim of udev's unpredictable interface naming scheme, which is on-by-default in recent versions of udev. As cwr says, you can tell it to use the traditional name by passing net.ifnames=0 on the kernel command line. You can also touch into existence an override file in /etc/udev/rules.d, but upstream has already renamed that file at least once, thus breaking anyone who relied on it.
With regard to your problem, when in the chroot, what is the output of cat -n /etc/resolv.conf ; ls -l /etc/resolv.conf?
cwr wrote: | enp3so is now considered a much better name than eth0. | That is a matter of opinion, and there are strongly held beliefs on both sides as to whether this on-by-default name mangling feature is a good idea. |
|
Back to top |
|
|
kurly Apprentice
Joined: 02 Apr 2012 Posts: 260
|
Posted: Wed Jun 04, 2014 1:54 am Post subject: Re: Chapter 6 errors |
|
|
Hu wrote: | cwr wrote: | enp3so is now considered a much better name than eth0. | That is a matter of opinion, and there are strongly held beliefs on both sides as to whether this on-by-default name mangling feature is a good idea. | Based on the rest of the post, I think you missed out on a bit of sarcasm. |
|
Back to top |
|
|
Albion n00b
Joined: 07 Feb 2012 Posts: 31 Location: England
|
Posted: Wed Jun 04, 2014 8:40 pm Post subject: |
|
|
Thanks for the help folks. I found I had indeed missed out copying the resolv.conf over. I've corrected that and also managed to configure Grub2 to use my monitors resolution (2560x1080). |
|
Back to top |
|
|
Pearlseattle Apprentice
Joined: 04 Oct 2007 Posts: 162 Location: Switzerland
|
Posted: Wed Jun 04, 2014 9:46 pm Post subject: Re: Chapter 6 errors |
|
|
kurly wrote: | Hu wrote: | cwr wrote: | enp3so is now considered a much better name than eth0. | That is a matter of opinion, and there are strongly held beliefs on both sides as to whether this on-by-default name mangling feature is a good idea. | Based on the rest of the post, I think you missed out on a bit of sarcasm. |
Is somewhere a website which explains what the new names are supposed to mean?
I'm still sticking with the old ones and I basically really hate the new ones but maybe it's because I'm not flexible - with a proper guide I might understand what the new names mean and assimilate + accept the idea.... .
Thx |
|
Back to top |
|
|
Albion n00b
Joined: 07 Feb 2012 Posts: 31 Location: England
|
Posted: Wed Jun 04, 2014 10:02 pm Post subject: Re: Chapter 6 errors |
|
|
Pearlseattle wrote: | kurly wrote: | Hu wrote: | cwr wrote: | enp3so is now considered a much better name than eth0. | That is a matter of opinion, and there are strongly held beliefs on both sides as to whether this on-by-default name mangling feature is a good idea. | Based on the rest of the post, I think you missed out on a bit of sarcasm. |
Is somewhere a website which explains what the new names are supposed to mean?
I'm still sticking with the old ones and I basically really hate the new ones but maybe it's because I'm not flexible - with a proper guide I might understand what the new names mean and assimilate + accept the idea.... .
Thx |
Hello there,
I don't know why the name changed. I'd have thought eth0 would be better than the longer, enp3s0 because it's easier (shorter) to type and to me, stands out much more readily on the screen or page. I've done a brief Google search but as yet not found the reason, just a lot of people complaining about it. I'll search more tomorrow though.
A. |
|
Back to top |
|
|
Hu Moderator
Joined: 06 Mar 2007 Posts: 21635
|
Posted: Thu Jun 05, 2014 2:15 am Post subject: |
|
|
The news item that originally announced this, Upgrading udev to version >=200 also includes a link to the upstream propaganda piece that touts why the new names are so much better. It also documents how the names are derived from card data, which may be useful if you need to determine what name it will assign to a mangled card. |
|
Back to top |
|
|
cwr Veteran
Joined: 17 Dec 2005 Posts: 1969
|
Posted: Thu Jun 05, 2014 2:39 pm Post subject: |
|
|
If you are running a large (or very large) server farm, with each blade (?) having multiple ethernet connections, then having explicit naming is a good idea. The large internet companies are thus all in favour, and having the money and the programmers they get their way.
For ordinary users it's a royal pain, since you never know what a port is called on a strange machine, or when hardware is updated. Few laptops or desktops have more than one ethernet or wireless port, so for ordinary users the naming of multiple ports is a non-problem.
Will |
|
Back to top |
|
|
Albion n00b
Joined: 07 Feb 2012 Posts: 31 Location: England
|
Posted: Thu Jun 05, 2014 2:40 pm Post subject: |
|
|
Thanks, Hu. |
|
Back to top |
|
|
depontius Advocate
Joined: 05 May 2004 Posts: 3509
|
Posted: Thu Jun 05, 2014 2:48 pm Post subject: |
|
|
There is only one network device name which is persistent and potentially predictable, and that name needs to be based on the MAC.
It is persistent because the MAC is. Even if you use macchanger, the network device's built-in MAC never changed. Using macchanger will just give you a network device whose effective MAC simply doesn't match the MAC embedded in the device name.
It is potentially predictable because most products shipping with network devices print the MAC somewhere on the box or on some paper somewhere in the box.
With network device names based on PCI enumeration, as they are now, plugging ANY pci/pcie device into the machine may change the enumeration, therefore changing the network device name.
It's even worse with USB network device names, because plugging USB devices is supposed to be commonplace. _________________ .sigs waste space and bandwidth |
|
Back to top |
|
|
Ant P. Watchman
Joined: 18 Apr 2009 Posts: 6920
|
Posted: Thu Jun 05, 2014 8:13 pm Post subject: Re: Chapter 6 errors |
|
|
Pearlseattle wrote: |
Is somewhere a website which explains what the new names are supposed to mean?
I'm still sticking with the old ones and I basically really hate the new ones but maybe it's because I'm not flexible - with a proper guide I might understand what the new names mean and assimilate + accept the idea.... .
Thx |
You're right to hate the new names. Whoever designed this scheme is smoking crack.
The new names represent which PCI/USB/PCMCIA/etc bus+port the network interface is in, where the numbers are just based on whatever order the kernel sees them in. This causes things like simple laptop wifi device hotplug or USB device startup delays cascading up the bus to effectively randomise your device names every time they appear. Those are real examples of problems people on f.g.o have had btw, they're not just made up.
Disable it completely, the only device name you can predict reliably (after chopping off udev's sticky tendrils) is eth0. |
|
Back to top |
|
|
Pearlseattle Apprentice
Joined: 04 Oct 2007 Posts: 162 Location: Switzerland
|
Posted: Thu Jun 05, 2014 11:40 pm Post subject: |
|
|
Hu wrote: | The news item that originally announced this, Upgrading udev to version >=200 also includes a link to the upstream propaganda piece that touts why the new names are so much better. It also documents how the names are derived from card data, which may be useful if you need to determine what name it will assign to a mangled card. |
Thanks guys for all the feedbacks! |
|
Back to top |
|
|
kernelOfTruth Watchman
Joined: 20 Dec 2005 Posts: 6111 Location: Vienna, Austria; Germany; hello world :)
|
Posted: Tue Jun 10, 2014 2:15 pm Post subject: Re: Chapter 6 errors |
|
|
Hu wrote: | Albion wrote: | Running ifconfig shows eth0 has disappeared and has been replaced with "enp3s0". I don't know what's going on because I've gone through the manual before and did a temporary installation of Gentoo which completed successfully. | Congratulations, you are a victim of udev's unpredictable interface naming scheme, which is on-by-default in recent versions of udev. As cwr says, you can tell it to use the traditional name by passing net.ifnames=0 on the kernel command line. You can also touch into existence an override file in /etc/udev/rules.d, but upstream has already renamed that file at least once, thus breaking anyone who relied on it.
With regard to your problem, when in the chroot, what is the output of cat -n /etc/resolv.conf ; ls -l /etc/resolv.conf?
cwr wrote: | enp3so is now considered a much better name than eth0. | That is a matter of opinion, and there are strongly held beliefs on both sides as to whether this on-by-default name mangling feature is a good idea. |
haha - me, too
had made a backup (stage4) and transferred to another computer
then been wondering why the heck dns wouldn't work anymore with NetworkManager (still haven't figured it out, yet - but this way it may be easier with net.ifnames=0)
thanks to all in this thread for posting the info for troubleshooting ! _________________ https://github.com/kernelOfTruth/ZFS-for-SystemRescueCD/tree/ZFS-for-SysRescCD-4.9.0
https://github.com/kernelOfTruth/pulseaudio-equalizer-ladspa
Hardcore Gentoo Linux user since 2004 |
|
Back to top |
|
|
Pearlseattle Apprentice
Joined: 04 Oct 2007 Posts: 162 Location: Switzerland
|
Posted: Tue Jun 10, 2014 6:08 pm Post subject: |
|
|
Hu wrote: | The news item that originally announced this, Upgrading udev to version >=200 also includes a link to the upstream propaganda piece that touts why the new names are so much better. It also documents how the names are derived from card data, which may be useful if you need to determine what name it will assign to a mangled card. |
Thank you Hu!
The first time I had a look at this I was overwhelmed, but now after 2 beers I was able to figure out the name of the interface before rebooting without "net.ifnames=0":
as "lspci" returned...
Quote: | # lspci | grep -i wireless
04:00.0 Network controller: Intel Corporation Wireless 7260 (rev 83) |
...I figured that the name would be:
- "wl" because it's wireless
- "p4" because the first digit of lspci is 04
- "s0" because the second digit of lspci is 00 (or maybe because the third one was 0 - not sure)
Result: wlp4s0, and it really matches the NIC's name after the reboot.
Not bad. It probably really does make things simplier if you do have a lot of NICs.
Cheers |
|
Back to top |
|
|
|