Ask your questions here.
Post a reply

Network problem with full persistence.

Fri Sep 06, 2013 12:02 am

Didn't think about this before, and in retrospect, it's obvious, but it had me confused for a little while yesterday.

1. Put refracta on a usb stick with a second partition for full persistence (i.e. all files, not just /home).
2. Boot the stick in my computer. All is well.
3. Take the stick to a friend's house and boot his computer with it. No wired network.
4. Kill wicd, edit /etc/network/interfaces for eth0 to use dhcp.
5. ifup etho, and get error message saying that eth0 does not exist.
6. Look in /etc/udev/rules.d/70-persistent-net and see a big list of network adapters, the one on the currently running computer is at the end of the list. eth0 is back at my house on another machine. Look in /etc/resolv.conf and see my nameserver and domain listed.
7. Delete those two files (in #6), remove the entry for eth0 from inerfaces file and reboot. Wired network works automatically with wicd, as it normally does.

Re: Network problem with full persistence.

Fri Sep 06, 2013 12:32 am

A live-config script can be made (old topic at fdn:)

http://forums.debian.net/viewtopic.php?f=17&t=61077

optical drives are probably also affected

Re: Network problem with full persistence.

Fri Sep 06, 2013 1:37 pm

Wow! Good memory. And nice to see you around again, dzz. Welcome back.

I just checked all my wheezy installs (all refracta installs) and the only files in /etc/udev/rules.d are the 70-persistent-net and 70-persistent-cd files. I thought there were supposed to be more files. In squeeze, there's also z60_hdparm.rules.

So, I'm wondering what the best way is to deal with this. And I don't understand the reason for disabling 75-persistent-net-generator.rules in your test in the linked post. Wouldn't that prevent udev from creating a new 70-persistent-net.rules, and isn't that file needed for the network to be functional?

Also, see your inbox. I'm about to send you a pm.

Re: Network problem with full persistence.

Sat Sep 07, 2013 11:04 am

Just back from a long summer break, the outdoor life beats indoors at the computer!
I don't understand the reason for disabling 75-persistent-net-generator.rules in your test in the linked post. Wouldn't that prevent udev from creating a new 70-persistent-net.rules, and isn't that file needed for the network to be functional?

The contents of /etc/udev/rules.d do not affect network functionality. They are a nuisance in a live session, as is manual reconfigs just to get a connection.

The only way I found to solve that problem using full persistence was prevent them being generated at all. Otherwise the network device becomes eth1 (or 2 or 3...) next boot in a different machine. Even if you delete those files and restart udev it's still eth1 (or 2 or 3...)

And if you use wicd (probably configured to use eth0 and wlan0) all you need in /etc/network/interfaces is:

Code:
auto lo
iface lo inet loopback

Doing this and rebooting one time only solved it for me:

Code:
mkdir /lib/udev/rules.d/disabled
mv /lib/udev/rules.d/75-persistent-net-generator.rules /lib/udev/rules.d/disabled
mv /lib/udev/rules.d/75-cd-aliases-generator.rules /lib/udev/rules.d/disabled
rm -f /etc/udev/rules.d/70*
Post a reply