network service failing on F23 base image

[Kushal Das]

Hi all,

For the last few days we can see that the *network* service is failing
on Fedora 23 Cloud base images [1].

Upon digging more in the issue, I found the following in the journal.

Oct 12 19:03:40 localhost.localdomain systemd[1]: Starting LSB: Bring up/down networking...
Oct 12 19:03:41 localhost.localdomain network[424]: Bringing up loopback interface:  [  OK  ]
Oct 12 19:03:51 testcloud.localdomain network[424]: Bringing up interface ens3:  ERROR    : [/etc/sysconfig/network-scripts/ifup-eth] Device ens3 does not seem to be present
Oct 12 19:03:51 testcloud.localdomain /etc/sysconfig/network-scripts/ifup-eth[680]: Device ens3 does not seem to be present, delaying initialization.
Oct 12 19:03:51 testcloud.localdomain network[424]: [FAILED]
Oct 12 19:03:51 testcloud.localdomain network[424]: Bringing up interface eth0:
Oct 12 19:03:51 testcloud.localdomain dhclient[702]: Created duid \000\001\000\001\035\256\300\227RT\000\036hd.
Oct 12 19:03:51 testcloud.localdomain dhclient[702]: DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 3 (xid=0x7e6c2257)
Oct 12 19:03:53 testcloud.localdomain dhclient[702]: DHCPREQUEST on eth0 to 255.255.255.255 port 67 (xid=0x7e6c2257)
Oct 12 19:03:53 testcloud.localdomain dhclient[702]: DHCPOFFER from 192.168.122.1
Oct 12 19:03:53 testcloud.localdomain dhclient[702]: DHCPACK from 192.168.122.1 (xid=0x7e6c2257)
Oct 12 19:03:55 testcloud.localdomain NET[730]: /usr/sbin/dhclient-script : updated /etc/resolv.conf
Oct 12 19:03:56 testcloud.localdomain dhclient[702]: bound to 192.168.122.83 -- renewal in 1756 seconds.
Oct 12 19:03:56 testcloud.localdomain network[424]: Determining IP information for eth0... done.
Oct 12 19:03:56 testcloud.localdomain network[424]: [  OK  ]
Oct 12 19:03:56 testcloud.localdomain systemd[1]: network.service: Control process exited, code=exited status=1
Oct 12 19:03:56 testcloud.localdomain systemd[1]: Failed to start LSB: Bring up/down networking.
Oct 12 19:03:56 testcloud.localdomain systemd[1]: network.service: Unit entered failed state.
Oct 12 19:03:56 testcloud.localdomain systemd[1]: network.service: Failed with result 'exit-code'.

While trying to find out a solution, I was pointed to [2]. I tried to follow
the initial solution given there and added * biosdevname=0 net.ifnames=0* on
the network section in the kickstart, and did a local build. But it still
failed with the same issue. So, the only *easy* solution I can think of
is actually deleting the file */etc/sysconfig/network-scripts/ifcfg-ens3* in the
post installation section of the kickstart file. If there is no other better
solution than this, I will go ahead and commit that for F23, and master.

[1] https://apps.fedoraproject.org/autocloud/jobs/255/output
[2] https://access.redhat.com/discussions/916973

Kushal

[Matthew Miller]

On Tue, Oct 13, 2015 at 05:23:00PM +0530, Kushal Das wrote:
> the network section in the kickstart, and did a local build. But it still
> failed with the same issue. So, the only *easy* solution I can think of
> is actually deleting the file */etc/sysconfig/network-scripts/ifcfg-ens3* in the
> post installation section of the kickstart file. If there is no other better
> solution than this, I will go ahead and commit that for F23, and master.

+1 -- sounds like the right thing to do.

[Dusty Mabe]

On 10/13/2015 08:38 AM, Matthew Miller wrote:
> On Tue, Oct 13, 2015 at 05:23:00PM +0530, Kushal Das wrote:
> > the network section in the kickstart, and did a local build. But it still
> > failed with the same issue. So, the only *easy* solution I can think of
> > is actually deleting the file */etc/sysconfig/network-scripts/ifcfg-ens3* in the
> > post installation section of the kickstart file. If there is no other better
> > solution than this, I will go ahead and commit that for F23, and master.
>
> +1 -- sounds like the right thing to do.

I am also interested in why this started showing up. Did we make a 
chance to our kickstarts? Was there a change made to the build servers?

Dusty

[Matthew Miller]

On Tue, Oct 13, 2015 at 08:44:08AM -0400, Dusty Mabe wrote:
> > +1 -- sounds like the right thing to do.
>
> I am also interested in why this started showing up. Did we make a
> chance to our kickstarts? Was there a change made to the build
> servers?

This sounds like an opportunity to tell my favorite tech joke.

Manager, hardware engineer, and software engineer are traveling to a
conference. At the top of a steep hill, brakes fail and all manner of
things go wrong and the car goes careening down the hill. Miraculously,
car gets to the bottom and everyone is unharmed.

Manager says: "That was scary but we made it. Let's keep going."

Hardware engineer: "No, no, we need to open the hood and figure out
what went wrong."

Software engineer: "Okay, but first, let's take it back to the top of
the hill and see if it happens again."