DevHeads.net

Postings by Dusty Mabe

Fedora 28 Atoimic Host RC 1.1 available for testing

Hash: SHA256

The Atomic Host compose based on RC 1.1 is available for testing now.

The toplevel directory is: <a href="https://kojipkgs.fedoraproject.org/compose/twoweek/Fedora-Atomic-28-20180425.0" title="https://kojipkgs.fedoraproject.org/compose/twoweek/Fedora-Atomic-28-20180425.0">https://kojipkgs.fedoraproject.org/compose/twoweek/Fedora-Atomic-28-2018...</a>

The cloud images are under the compose/AtomicHost/$arch/images/ directories.
The ISO images are under the compose/AtomicHost/$arch/iso/ directories.

We don't yet have AMIs. Will update you when those are available.

- - Atomic WG

march VFAD for the atomic working group

Hash: SHA256

<a href="https://pagure.io/atomic-wg/issue/429" title="https://pagure.io/atomic-wg/issue/429">https://pagure.io/atomic-wg/issue/429</a>

In the Atomic working group in Fedora we have occasionally used virtual
Fedora Activity Days (VFADs) to get the community together to discuss
in a high bandwidth setting (video conference) some techical issues we're
having or some direction of where we're going.

We have scheduled a VFAD for next Friday (march 9th). We'd like to get
some input on some topics to discuss during this VFAD.

runroot changing during the course of a rawhide pungi run

Hash: SHA256

If I understand correctly there was a new lorax build [1] that completed
at 01:11 UTC (02/23) that then made it into the runroot of a task [2] that
was part of a pungi compose [3] that started at 14:54 UTC (02/22).

I think this caused problems with the build, but that isn't really
important. The real question is: if our runroot changes during the run then
we could have some tasks (lorax imagebuild etc) that run with some versions
of software and other tasks that run with others.

Atomic Host/Workstation Interview for Linux Unplugged Podcast

Hash: SHA256

A couple of us went on the LUP podcast this week to talk about Atomic Host/Workstation.

Check it out: <a href="https://twitter.com/jupitersignal/status/961283877883977728" title="https://twitter.com/jupitersignal/status/961283877883977728">https://twitter.com/jupitersignal/status/961283877883977728</a>. We start around minute 16:20.

Dusty

make yum repo when we build things in koji

Hash: SHA256

I often build an rpm in koji to make sure it builds but in order to
consume that rpm using our tools most of the time it needs to be in
a yum repo somewhere. I could create a repo locally and serve it locally
or even push it to somehwere to fedorapeople. The problem is, I'm lazy.
I'll often just take the srpm from koji and import it into copr and then
consume that repo once the build has finished over there (thanks copr)

The problem with this is that it's kind of a waste of resources.

please help us test the f27 cloud images in RC1.2

You can grab the image download links and fill out the results here:

<a href="https://fedoraproject.org/wiki/Test_Results:Fedora_27_RC_1.2_Cloud" title="https://fedoraproject.org/wiki/Test_Results:Fedora_27_RC_1.2_Cloud">https://fedoraproject.org/wiki/Test_Results:Fedora_27_RC_1.2_Cloud</a>

Unfortunately only half the AMIs uploaded:

Fedora-Cloud-Base-27-1.2.x86_64 EC2 (us-east-1) ami-acaa02d6 paravirtual gp2
Fedora-Cloud-Base-27-1.2.x86_64 EC2 (us-east-1) ami-00ac047a paravirtual standard
Fedora-Cloud-Base-27-1.2.x86_64 EC2 (us-west-2) ami-38b07b40 hvm standard
Fedora-Cloud-Base-27-1.2.x86_64 EC2 (us-west-1) ami-695867

Fedora Atomic WG - Changes to IRC/Mailing list

TL;DR There is a new fedora atomic mailing list [1] and IRC channel [2], but the mailing list
should only be used for automated messages and the IRC channel should be used only
when something extremely Fedora specific (like releng bits) is being discussed.

bex represents pagure emails in hyperkitty

The atomic working group has configured pagure to send notifications to the
cloud mailing list. These notifications come through as normal emails
but for some reason in hyperkitty the emails from pagure show up as
coming from Brian Exelbierd (bex).

new gemspec behavior in rawhide

Hey all,

In my vagrant-sshfs rpm i call `gem spec` here [0]. In rawhide vs f25 I get differing
outputs. Basically in rawhide I get a bunch of `.freeze` added to my strings like:

s.authors = ["Dusty Mabe".freeze]

See the full diff at [1]. This is causing pain for me because I actually patch the
gemspec file. Was this a desired change?

cloud.centos.org is down

I first noticed when trying to pull vagrant images.

<a href="http://downforeveryoneorjustme.com/cloud.centos.org" title="http://downforeveryoneorjustme.com/cloud.centos.org">http://downforeveryoneorjustme.com/cloud.centos.org</a>

Dusty

fc24 packages when running dnf on f25

Is it normal to have fc24 packages get installed when I install a package on F25?
This system started as an F25 cloud instance based on the cloud image.

$ rpm -qa | grep fc24 | wc -l
116
$ cat /etc/redhat-release
Fedora release 25 (Twenty Five)

- Dusty

unretire kompose package

I am trying to package [1] in fedora and there is a name conflict with an old/retired
package [2] that no longer has an upstream. I have talked with the old maintainer (cc'd)
and we would like to unretire the package and allow for the new project to assume the name
in rpm.

I'm attempting to follow the process described at [3].

t460s fedora 24 - display freeze

Several times over the past month my t460s has just blank screens
when I come back to it after some time. I don't know if the thing
froze up completely or if it is just the display not working. I lock
my screens (laptop + 2 external monitors) and then leave for an hour
or so. When I come back no attempt at input will get the lockscreen
to appear.

Broken: Firefox 48 + Private Tab + Kerberos SSO

I can't seem to get firefox-48.0-5.fc24.x86_64 to work with kerberos
single sign on in a private window. It works fine when using a
non-private window.

Any ideas on why this would have broken? Anyone else seeing this?

Dusty

iwl8000 driver in Fedora 24

I have the Lenovo T460s with the Intel Wireless AC 8260 chip in it.
According to the page at [1], that chip is supported by the
iwlwifi-8000 firmware. I don't see that package as an option in the
Fedora 24 repos [2]. The newest I see is the 7260. Googling around
hasn't led me to a solution. Anybody else having this same problem?

Thanks,
Dusty

[1] <a href="http://www.intel.com/content/www/us/en/support/network-and-i-o/wireless-networking/000005511.html" title="http://www.intel.com/content/www/us/en/support/network-and-i-o/wireless-networking/000005511.html">http://www.intel.com/content/www/us/en/support/network-and-i-o/wireless-...</a>
[2] <a href="http://mirrors.rit.edu/fedora/fedora/linux/updates/24/x86_64/i/" title="http://mirrors.rit.edu/fedora/fedora/linux/updates/24/x86_64/i/">http://mirrors.rit.edu/fedora/fedora/linux/updates/24/x86_64/i/</a>

T460s + latest kernel doesn't boot

kernel-4.6.4-301.fc24.x86_64

Once I select the latest kernel in grub it then pops up and says

Probing EDD (edd=off to disable)... OK

and then nothing else and stays there.

Anyone else having this problem? If i boot the original kernel from f24
(kernel-4.5.5-300.fc24.x86_64) then the system boots fine.

Any tips?

Thanks,
Dusty

iphone mount in nautilus on Fedora 24

In older Fedora (at least Fedora 22) I could easily mount my ihphone
in Nautilus. I could click on it in the left hand side of the File
Manager and it would mount.

I am having trouble getting this same behavior in Fedora 24. Can
someone verify this works or does not work for them? You should be
able to verify this with Fedora 24 + Gnome + iphone + iphone cable.

Thanks,
Dusty

centos-ci - new kube version broken

The version is: kubernetes-client.x86_64 0:1.2.0-0.6.alpha1.git8632732.el7

The error looks something like:

```
1458139640 - [ERROR] - cli/main.py - cmd: ['/host/usr/bin/kubectl', 'create', '-f', u'build/artifacts/kubernetes/.hello-apache-pod.json', u'--namespace=default'] failed:

Error from server: error when creating "build/artifacts/kubernetes/.hello-apache-pod.json": pods "helloapache" is forbidden: service account default/default was not found, retry after the service account is created
```

A full log is at [1].

Making mailman3/hyperkitty send my emails to me

I have played with all of the settings in hyperkitty and I still am
failing to get the emails that I compose and send to a mailing list
also sent back to me.

The reason I want this is because of the "archived-at" email header.
Once I compose an email and get it sent to me from mailman it will
have that header in the email and I'll be able to copy/paste a
permalink to it easily (see [1]).

Anyone experiencing the same issue?

Thanks,
Dusty

[1]
<a href="http://dustymabe.com/2016/01/10/archived-at-email-header-from-mailman-3-lists/" title="http://dustymabe.com/2016/01/10/archived-at-email-header-from-mailman-3-lists/">http://dustymabe.com/2016/01/10/archived-at-email-header-from-mailman-3-...</a>

Disabling predictable interface names

According to [1] predictable interface naming can be disabled with a
symlink to /dev/null or by providing net.ifnames=0 on the kernel command
line. It seems like symlinking to /dev/null isn't working any more.

Is symlinking to /dev/null supposed to still work? We are seeing this on
Atomic Cloud images where the symlink is in place but net.ifnames=0
isn't on the cmd line so we are getting ens* interface names.

[1] -
<a href="http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames/" title="http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames/">http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInter...</a>

Fedora Rel-Eng builder boxes can't run qemu-kvm on F23

Basically this means we can't build any images (rawhide/Fedora23/etc)
because the builder hosts have been moved to Fedora 23. The following BZ
has more info. Please discuss in the bug if you have comments about the
technicals. Please discuss here otherwise.

<a href="https://bugzilla.redhat.com/show_bug.cgi?id=1283666" title="https://bugzilla.redhat.com/show_bug.cgi?id=1283666">https://bugzilla.redhat.com/show_bug.cgi?id=1283666</a>

Canceled Atomic SIG meeting today

A number of us will be unavailable for the meeting today so we have
elected to cancel. See you next week!

Dusty

ansible in Fedora 23+ (python3)

Hey all,

Just wondering if anyone has started to feel any pain from trying to
use ansible with F23 systems. As part of the Fedora Cloud working
group it would be nice to ship a system that could be targeted by
ansible out of the box. Unfortunately, since ansible uses python 2.X
vs python 3 it doesn't really work.

A simple answer would be to "just install python", but I actually hit
several roadblocks because of python packages that are usually there
but have been replaced by their python 3 counterparts.

20151001 Atomic seems to be broken

For some reason I can't seem to get it [1] to boot. I get a "No
bootable device" massage from the BIOS for both libvirt/kvm and a
local openstack setup. Can someone confirm or refute this?

Dusty

[1] - <a href="http://cloud.centos.org/centos/7/atomic/images/CentOS-Atomic-Host-7.20151001-GenericCloud.qcow2" title="http://cloud.centos.org/centos/7/atomic/images/CentOS-Atomic-Host-7.20151001-GenericCloud.qcow2">http://cloud.centos.org/centos/7/atomic/images/CentOS-Atomic-Host-7.2015...</a>

CentOS/Fedora Atomic Test Day 09/22

Hey all,

Sorry for the wide distribution.. We'd like to promote the joint test
day we are doing next week for Atomic. Please see Jason's blog post for
more details and join us on the test day:

<a href="http://www.projectatomic.io/blog/2015/09/combined-fedora-centos-atomic-test-day-next-week/" title="http://www.projectatomic.io/blog/2015/09/combined-fedora-centos-atomic-test-day-next-week/">http://www.projectatomic.io/blog/2015/09/combined-fedora-centos-atomic-t...</a>

- Dusty

Request to join CentOS Atomic SIG

Hey Everyone,

As a member of the Fedora Cloud Working Group (the group within Fedora
that is most focused on atomic) I am interested in joining the CentOS
Atomic SIG. My hopes are to find differences in the Fedora and CentOS
versions of Atomic and try to minimize those where it makes sense so
that we can provide a more consistent approach for people using
Atomic Host. Additionally it would be nice to find places where Fedora
and CentOS can re-use the same or similar processes to generate these
deliverables.

Atomic SIG: Joint Test Day with Fedora Atomic Folks

Hi All,

There has been some discussion of doing a joint test day with Fedora
and CentOS atomic folks. Fedora is having a test day for Fedora Cloud
base on 9/8 (all are welcome to join us). We were thinking about
having the joint Fedora/CentOS test day on the following Tuesday
(9/15). Is this something everyone is interested in?

To do the test day it would be worth having a test day page like [1]
and some test cases for people to execute.

docker 1.8 issues with pulling images

I was trying to use docker 1.8 [1] and I keep having issues pulling from the
registry. See the issue upstream at [2]. Does anyone else see this?

Dusty

[1] - <a href="http://cbs.centos.org/koji/buildinfo?buildID=3714" title="http://cbs.centos.org/koji/buildinfo?buildID=3714">http://cbs.centos.org/koji/buildinfo?buildID=3714</a>
[2] - <a href="https://github.com/docker/docker/issues/15582" title="https://github.com/docker/docker/issues/15582">https://github.com/docker/docker/issues/15582</a>

cloud-init / dnf selinux denials

Hey all,

We have a bug that affects booting cloud images (see [1]). There are
denials that don't allow cloud-init to install packages. I'm not
exactly where a fix needs to be applied but I have copied both the
owners of dnf and the owner of selinux-policy on this email. What can
we do to get this looked at and fixed?

Thanks,
Dusty

<a href="https://bugzilla.redhat.com/show_bug.cgi?id=1227484" title="https://bugzilla.redhat.com/show_bug.cgi?id=1227484">https://bugzilla.redhat.com/show_bug.cgi?id=1227484</a>