Announcement

**SteveRiley** · Jun 14, 2015, 01:48 PM

It's obvious that the NetworkManager maintainer is unaware of how *buntu implements DNS and DHCP using Dnsmasq and Resolvconf. Please ignore the statement about a "rarely used" configuration and do not follow the advice to remove Dnsmasq.

What's the output of

Code:

cat /etc/network/interfaces

**gnomek** · Jun 14, 2015, 03:15 PM

When connection works it is:

Code:

# interfaces(5) file used by ifup(8) and ifdown(8)
auto lo
iface lo inet loopback

But I can't reproduce this "lack of connection" situation, so to give output I will have to wait until it happens.

**SteveRiley** · Jun 14, 2015, 03:24 PM

The file is correct.

When the problem happens, please open a console window and run:

Code:

cp /var/log/syslog ~/syslog.save

I'm having you do this to copy the log into a new file. Then reboot your computer. Once you get a connection, do this:

Code:

sudo apt-get install pastebinit

Now run:

Code:

pastebinit ~/syslog.save

The pastebinit utility will copy the saved file to Ubuntu's pastebin and return a URL. Reply here with that URL, so we can take a look at the log.

**Qqmike** · Jun 14, 2015, 03:40 PM

As you keep working on the software diagnosis as S-R advises ... Interjection: This is not your Ethernet cable, right? Or, even if the cable may be good, I'd be curious to see this: you log in, it doesn't work this time, with the PC booted as is, try removing the Ethernet cable at one end (either the modem end or the PC end where it plugs into the back of your PC), wait say 30 seconds, and plug it back in, see if the connection starts. I have this problem in 14.04 on some unpredictable schedule that seems random and may be once every month or two, or 4x a week.

**gnomek** · Jun 17, 2015, 02:30 AM

When I was disconnected

Code:

cat /etc/network/interfaces
# interfaces(5) file used by ifup(8) and ifdown(8)
auto lo
iface lo inet loopback

http://paste.ubuntu.com/11729368/

Unplugging Ethernet cable and putting it back started connection but this is not good as a default solution.
The biggest problem is that in this version of network manager reconnect option (green round arrow icon) doesn't work.
Do you think I should report a bug on Kubuntu launchpad and ask for updating network manager to the new version?

**Qqmike** · Jun 17, 2015, 05:40 AM

Unplugging Ethernet cable and putting it back started connection but this is not good as a default solution.

Oh I do agree with that. I'll be watching now to see if anything can be gleaned as to what could cause this. Thanks.

**SteveRiley** · Jun 17, 2015, 10:12 PM

Originally posted by gnomek View Post

When I was disconnected

Code:

cat /etc/network/interfaces
# interfaces(5) file used by ifup(8) and ifdown(8)
auto lo
iface lo inet loopback

http://paste.ubuntu.com/11729368/

Unplugging Ethernet cable and putting it back started connection but this is not good as a default solution.

Here's what I see.

1. On 6/15 at 10:54:00, NetworkManager noticed that a connection went away and thus deconfigured all networking.

2. Still at 10:54:00, DBus (a system-wide interprocess communication mechanism) reactived NetworkManager.

3. Between 10:54:32 and 10:54:34, NetworkManager configured eth0 for IPv4. DHCP asked for and obtained the address 192.168.0.10, gateway 192.168.0.1, and DNS servers 62.179.1.{60,61}. Using DBus, NetworkManager sent DNS details to dnsmasq, which is acting as your machine's local resolver.

4. At 10:54:52, NetworkManager gave up trying to configure the interface for IPv6 -- totally normal if you don't have IPv6 in your home network.

5. At 12:28:28, you rebooted your computer.

6. Between 12:28:31 and 12:28:36, NetworkManager configured eth0 for IPv4. DHCP asked for and obtained the address 192.168.0.10, gateway 192.168.0.1, and DNS servers 62.179.1.{60,61}. NetworkManager started dnsmasq (because it wasn't running) and tried to communicate with it, which failed -- dnsmasq wasn't ready yet. When dnsmasq appeared on DBus, NetworkManager sent DNS details to dnsmasq.

7. At 12:28:51, NetworkManager gave up trying to configure the interface for IPv6.

This all looks rather normal, but I have two questions.

What were you doing on your computer at 10:54:00 that might have caused the network to reconfigure itself? Was this when you unplugged and re-plugged the Ethernet cable?
Why did you reboot at 12:28:28?

Originally posted by gnomek View Post

The biggest problem is that in this version of network manager reconnect option (green round arrow icon) doesn't work.
Do you think I should report a bug on Kubuntu launchpad and ask for updating network manager to the new version?

I don't think your problem is related to the installed versions of NetworkManager or Plasma-NM. (Besides, 14.04 will not receive updates to these packages. *buntus aren't rolling releases.)

**gnomek** · Jun 18, 2015, 02:20 AM

I don't remember what I was doing.

Strange is that I created report before sending Wed, 17 Jun 2015, but in report itself there is Jun 15 listed.

Next time I will notice details and send another report.

**jlittle** · Jun 20, 2015, 10:53 PM

I'm not sure this is relevant, but it is similar to a long-standing problem involving Network Manager on waking up from suspension. It was supposedly fixed some time ago but still affects me occasionally on Utopic. Recently I researched it and found the command

sudo nmcli nm sleep false

It worked the only time I I've had to try it. (I had tried disconnecting and reconnecting the cable, and it didn't work, though perhaps I didn't wait long enough.)

**gnomek** · Jun 23, 2015, 03:39 AM

It happened again. I used

Code:

pastebinit ~/syslog.save

but it doesn't work properly. It displays old logs from Jun 15, so I don't link it here.

Instead here is today's log.
http://pastebin.com/Fwr7J9Jv
I turned PC on Jun 23 11:12:46
Relogged to administrator account to see logs but there was no connection to post them so next
Jun 23 11:17:48 - I restarted system. Connection started.

So, relogged again to administration account to get logs again and relogged to user account to post them.

**ripper81** · Jun 23, 2015, 02:34 PM

Hey same problem here.

I am running Kubuntu 15.04. I have two PC's with the exact same problem. One of those received a fresh installation, but the problem presists.
I have never experienced something like this with older Versions of Kubuntu. It also seems that it only effects hardwired connections.

Greetings ripper81

**pierpiotr** · Jun 25, 2015, 02:32 AM

Same pb here, almost

Hello friends in troubleland,
I have 15.04 at the office* and every morning I need to run

Code:

[FONT=monospace][COLOR=#000000]sudo service network-manager restart[/COLOR][/FONT]

I realised today that I can do it from the GUI. Nevertheless, it is quite frustrating to have such a nice OS and having to manually tweak something that should be automatic.
I noticed also that a reboot or restart the same day makes it unecessary to restart the network, if that helps.
here is

Code:

pier@pier-Bur:~$ cat /etc/network/interfaces

 # interfaces(5) file used by ifup(8) and ifdown(8)
 auto lo
 iface lo inet loopback

and pastebinit today
It would be lovely to move towards a solution to this.

*(15.04 at home as well, my signature is obsolete and I have not seen how to edit it)

**SteveRiley** · Jun 25, 2015, 11:08 PM

Originally posted by gnomek View Post

It happened again. I used

Code:

pastebinit ~/syslog.save

but it doesn't work properly. It displays old logs from Jun 15, so I don't link it here.

That's because the file syslog.save still contains what you copied into it the first time. But since you know how to successfully send syslogs to Pastebin, we don't need to use this procedure anymore.

Originally posted by gnomek View Post

Instead here is today's log.
http://pastebin.com/Fwr7J9Jv
I turned PC on Jun 23 11:12:46
Relogged to administrator account to see logs but there was no connection to post them so next
Jun 23 11:17:48 - I restarted system. Connection started.

So, relogged again to administration account to get logs again and relogged to user account to post them.

In your PM to me, you wrote:

I wrote to NetworkManager developer and he replied:

"When I wrote that using 127.0.0.1 as nameserver is rarely used I meant rarely used by normal users. Usually only servers use such a configuration, if *Ubuntu decided to use it then you are susceptible to problems related to the interaction between NetworkManager and the nameserver (dnsmasq server in this case). I prefer not to use dnsmasq for something as simple as updating /etc/resolv.conf, it may add more flexibility but also adds more things to go wrong. You can try configuring your system to launch dnsmasq before NetworkManager starts, that way the first NetworkManager's attempt to update /etc/resolv.conf should not fail."

I don't know anything about config file configuration in connection to internet.

Do you think it is reasonable to report Kubuntu bug at launchpad?

I don't believe your problem is related to how Ubuntu configures the network stack. If there really were an issue related to how NetworkManager, dnsmasq, and resolvconf are configured, then hundreds of thousands of people would all experience the same problem. In Ubuntu's configuration, resolv.conf is actually managed by the package resolvconf. dnsmasq receives DNS information from DHCP and stores it in DBus. dnsmasq resolves queries. resolvconf ensures that the resolv.conf file always points to the local dnsmaq instance rather than a DNS server somewhere on the Internet. This becomes especially important when you're using a VPN. dnsmasq knows which name servers can resolve addresses for which subnets, and insures that you can always type only a hostname (as opposed to a fully qualified domain name) and it will be resolved on the correct network. This configuration has been the default *bunutu way for many releases now, is quite robust, and shouldn't be changed. In fact, other distros could learn from it.

As I look at your logs, I see an important difference. During the 11:17 attempt, we see this:

Code:

Jun 23 11:17:52 yellow-945P-DS3 NetworkManager[810]: <info> (eth0): carrier now ON (device state 20)

That never appeared for th 11:12 event. For some reason, the Ethernet jack isn't fully enabling itself after a cold boot.

When the network is working correctly, what's the output of

Code:

sudo lshw -C network

Do you have any sort of power management software running?

**gnomek** · Jun 26, 2015, 03:24 AM

Code:

  *-network               
       description: Ethernet interface
       product: RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller
       vendor: Realtek Semiconductor Co., Ltd.
       physical id: 0
       bus info: pci@0000:03:00.0
       logical name: eth0
       version: 01
       serial: **:**:**:**:**:**
       size: 1Gbit/s
       capacity: 1Gbit/s
       width: 64 bits
       clock: 33MHz
       capabilities: pm vpd msi pciexpress bus_master cap_list rom ethernet physical tp mii 10bt 10bt-fd 100bt 100bt-fd 1000bt 1000bt-fd autonegotiation
       configuration: autonegotiation=on broadcast=yes driver=r8169 driverversion=2.3LK-NAPI duplex=full ip=192.168.0.10 latency=0 link=yes multicast=yes port=MII speed=1Gbit/s
       resources: irq:43 ioport:a000(size=256) memory:d4000000-d4000fff memory:80400000-8040ffff

Originally posted by SteveRiley View Post

Do you have any sort of power management software running?

No. And I don't have any special settings set. It is classic PC, not laptop.

Announcement

network connection doesn't start

network connection doesn't start

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment