Installing fedora in a new (trash can) mac pro
by Rafael Espíndola
I recently managed to install fedora in a trash can mac pro. I am not
sure what the best place to document the steps, so I am just sending
an email. Hopefully search engines can find it.
Let me know if you think something should be reported as a bug.
First, a list of the problems:
* The kernel that comes with the installer will not boot. I am not
sure what the issues is, but yum update will fetch a kernel that
works. Unfortunately I could not find a way to rebuild the
installation image from sources.
* Using glub2 is just too horrible for me. It is incredibly slow. At
first I also thought it might be responsible for the kernel not
booting, but that was not the case. The good news is that gummiboot
works *really* well.
* The apple firmware is very slow at booting from the FAT ESP
partition. It is way faster at loading anything from hfs.
* The linux hfs driver will not write to a hfs partition that has
journalling enabled.
* To get the ui running I had to pass nomodeset to the kernel.
* If the monitor (a dell dell up2414q) is connected vi display port,
it will not come back from sleep. It is slow to wake up on OS X, but
works. Connecting it with an hdmi cable avoids the problem.
The steps to get the installation working:
First, create a VM:
* Using virtualbox create a VM which uses EFI and install fedora in it.
* Use yum update to get a new kernel.
* Install gummiboot, remove grub2. Make sure you can still boot the VM.
* You can now delete the /boot partition and make /boot just a
directory. Gummitboot will be using /boot/efi
* Boot the vm with the install cd again and create a tar file of the
root fs and copy it to the host. I used tar --selinux --acls --xattrs
... | nc
Now, create the partitions for fedora;
* Using disk util reduce the size of the os x install partition.
* Create a small partition for gummiboot to live in, make it hfs.
* Disable journalling in the gummiboot partition.
Install gummiboot:
* Copy it to the new hfs partition
* run bless:
sudo bless --setBoot --folder /volumes/gummiboot --file /Volumes/gummiboot.efi
* Rebooting the system should show up gummiboot with only one entry: os x.
Copying fedora to the new partitions:
* Download a live cd with a new kernel. I used ubuntu 14.04. It
requires nomodeset to start the UI.
* With the livecd, create a btfs partition in the free space after the
gummiboot partition.
* Use tar -xvpf as root to extract the root contents.
* Update /etc/fstab to have the new partition names and UUIDs.
* Copy the kernel and initrd to the gummiboot partition and create an
entry in the config (remember to include nomodeset).
* Reboot and pass selinux=0 to the kernel
* Login as root.
* touch /.autorelabel
* reboot.
I am sure there are more efficient ways of doing this. In particular,
the relabelling is probably unnecessary if you use a live cd with
selinux.
Was it worth it? Yes, linux is quiet a bit faster. Running the llvm
and clang tests now takes 32.18s, under OS X it takes 49.56s.
Cheers,
Rafael
9 years, 11 months
can't find skype for Fedora 20
by Robert Moskowitz
I have rpmfusion-free and rpmfusion-nonfree repos set up. When I do a
yum whatprovides skype
I get No matches found.
9 years, 11 months
About NETWORK CONFIGURATION in kickstart process
by Noriaki TAKAMIYA
Hi,
I'm new to anaconda, but I have one question.
I'm going to install Fedora 20 with DVD media and kickstart file to
PC which has no network interfaces.
Everything is going well except that kickstart process stops at the
NETWORK CONFIGURATION window with the message "We'll need network
access to fetch information about your location and to make software
updates available for you".
Could you give me how to skip this NETWORK CONFIGURATION window?
Regards,
P.S. I attach using kickstart.cfg
-----
#version=DEVEL
# System authorization information
auth --enableshadow --passalgo=sha512
# Use CDROM installation media
cdrom
# Run the Setup Agent on first boot
firstboot --enable
# Keyboard layouts
keyboard --vckeymap=us --xlayouts='us'
# System language
# lang ja_JP.UTF-8
# Network information
# network --hostname=localhost.localdomain
#network --bootproto=static --device=lo --ip=127.0.0.1
--netmask==255.0.0.0 --nodefroute --nodns --hostname=localhost.lcoaldomain
# Root password
rootpw --iscrypted
$6$etxGu6z/bVEGULjV$PTypYz.NpPNmOLxgreym7PbdJHmE9evT7RRsfQY8JNsLNbXLmhT.Cmq9KjKD8c98SowkRN5MZjtEEshz03rCV1
# System timezone
timezone --utc Asia/Tokyo
# System bootloader configuration
autopart --type=plain
# Partition clearing information
#clearpart --none --initlabel
clearpart --all --drives=vda
bootloader --location=mbr --boot-drive=vda
ignoredisk --only-use=vda
%packages
@core
%end
--
NT
9 years, 11 months
configure trackpad
by CS DBA
All
I have a thinkpad W540 running Fedora 20, the trackpad seems to have
sort of physical left/right buttons meaning the entire pad sort of
'moves' when i click and it does a right click if i click on the right
side of the trackpad (single finger click) and a left click on the left
side. It drives me crazy since I'm used to a single finger click doing
a left mouse click no matter where I click and a right click if I do a
two finger click.
The trackpad still does a right click if I use 2 fingers, how can I
configure it to do a left click when I use a single finger no matter
where on the trackpad I click?
Thanks in advance
9 years, 11 months
CUPS service does not running F18
by David Carpio
Hello everybody:
A month ago I have the same problem when Fedora 18 is running always
shows me this message:
Process /usr/cupsd was killed by signal 11 (SIGSEGV)
And I cannot enter to CUPS interface
http://127.0.0.1:631
I have reinstalled again and again but it is the same result.
Any suggestions gratefully received.
David
9 years, 11 months
Power Management, SystemD, etc.
by Trever L. Adams
Hello everyone,
I am having some difficulty finding the right information. I have a
server that is only needed at certain times. Wake-on-lan works well, so
I can do administration, etc.
rtcwake works, so I can tell it to wake each day 3-5 minutes or so
(haven't decided) before when it is needed.
The problem I am having is figuring out how to set it up so it
automatically goes to sleep after X period of time and stay asleep.
This is a F20 system, so systemd is installed. I would like to have a
script get called prior to sleep so it can set rtcwake.
I cannot find any documentation for a non-X/gui/gnome/whatever system to
set up the idle time out. Do the hooks in /etc/pm/sleep.d work with systemd?
I prefer to do suspend to ram for this setup.
Thank you for any help,
Trever
9 years, 11 months
Five Things in Fedora This Week (2014-05-27)
by Matthew Miller
Reposted from
http://fedoramagazine.org/five-things-in-fedora-this-week-2014-05-27/
Fedora is a big project, and it’s hard to follow it all. This series
highlights interesting happenings in five different areas every week.
It isn’t comprehensive news coverage — just quick summaries with links
to each. Here are the five things for May 27th, 2014:
Flock Schedule!
---------------
Flock is Fedora’s annual contributor conference, held this year in
Prague on August 6th through 9th. The schedule is now up on sched.org.
The organizers reserve the right to make tweaks as needed, but this
should mostly be the final schedule.
Last year’s conference was amazing, and I think this year will be even
better. There are a lot of great talks, hackfests, and workshops
planned over a busy four days, with a vibrant “hallway track” and fun
social events (yet to be detailed). In general, we will get a lot of
real planning and solid work done, all built on top of Fedora’s
“Friendship” foundation. If you’re a Fedora contributor and not sure if
you should come, *you should*. Register now at the Flock site.
Thanks to everyone who proposed talks and to everyone who voted on
those proposals. And thanks to Flock organizers Josh Boyer, Tom
Callaway, and Ruth Suehle, who used those votes as the primary source
for the difficult jigsaw puzzle that is putting together the schedule.
Thanks also to Miro Hrončok and Jiri Eischmann for working out venue
logistics.
* http://flocktofedora.org/
* http://flock2014.sched.org/grid/
Fedora 21 Schedule Reminder
---------------------------
Speaking of schedules… as we get to the end of May, it seems like a
good time to give a reminder of the Fedora 21 schedule. The summer is
going to go fast! So, a refresher:
- **June 6:** *Mass Rebuild.* This means that Release Engineering
will cause every package in the distribution to be rebuilt from
source, with the newest compilers and toolchains. Package
maintainers need to be prepared to fix any problems that are
exposed. (Packages that don’t build and aren’t updated will
eventually be dropped from the distribution.)
- **July 8:** *Change Freeze* and *Branch from Rawhide*. The first of
these means that any official change proposals for F21 need to be
a) substantially complete and in a testable state and b) enabled by
default (if that’s the plan for that change). We have a rather
large change set planned for this release, and many of those
changes involve big new Fedora.next ideas. The second means that
Fedora 21 will become a distinct repository from Rawhide, our
rolling development tree. That means that any changes packagers
make that are intended for F21 will go through the normal package
updates process — and it means that testers can decide if they want
to divert their systems to the upcoming release, or to keep on
Rawhide to follow the bleeding edge of what will eventually become
the release *after* next.
- **July 22:** *Alpha Change Deadline* and *Software String Freeze*.
The change deadline basically means that except for accepted fixes,
everything that’s going to land in the alpha release should be
there already. And the string freeze means that packages for which
Fedora is the upstream should stop making user-visible changes so
that translators have time to work.
- **August 5:** *Alpha Release*. The Fedora 21 Dress Rehearsal — we’ll
get real picture of what the upcoming release will look like (and
- **August 26:** *Beta Change Deadline*, *Changes 100% Complete*, and
*Software Translation Deadline*. Similar to the deadlines for
Alpha, but more so. Proposed changes which are off-track at this
point may need to be scaled back or delayed until the next release.
- **September 9:** *Beta Release*. Everything will be shaping up, and
we’ll be more concerned with polish than development. If you have
an idea for a big change at this point, that’s awesome — the
Rawhide development tree continues to be open for business so
progress can continue.
- **September 30:** *Final Change Deadline*. All the last changes
should be in.
- **October 14:** ***Fedora 21 Final Release!*** Fedora Cloud, Fedora
Server, Fedora Workstation, and all the spins and the rest will go
into the world.
All of these dates are officially expressed as “no earlier than”. As
always the schedule may… flex… a little bit, but generally we aim to
actually hit these targets. That gives us just 20 weeks from now — mark
your calendars!
* http://fedoraproject.org/wiki/Releases/21/Schedule
* https://fedoraproject.org/wiki/Releases/21/ChangeSet
* http://fedoraproject.org/wiki/Package_update_HOWTO
* https://fedoraproject.org/wiki/L10N
Fedora.next Preliminary Branding
--------------------------------
The Fedora Design Team is working on a plan for presentation of the
three different Fedora products we’ll be releasing this fall as part of
the Fedora.next initiative. Each of these will need their own logo, but
they also need a coherent look that ties them together as part of the
overall Fedora Project. Fedora Contributor Máirín Duffy kicks off the
process with some starting-point designs, which you can see at read
about at her blog.
What do you think? What ideas do you have?
* https://fedoraproject.org/wiki/Design
* http://fedoramagazine.org/fedora-present-and-future-a-fedora-next-2014-up...
* http://blog.linuxgrrl.com/2014/05/22/fedora-next-brand-concept-1/
Fedora 21 Test Planning
-----------------------
Those three products will also need testing; that’s one of the reasons
the schedule is longer than the typical six months. Fedora Quality
Assurance “Community Monkey” Adam Williamson wrote a draft test plan
for Fedora 21 in general, outlining different areas of work and
responsibilities. If you’re interested in helping (or just curious),
read Adam’s message about this draft and join the conversation on the
Fedora Test Mailing List.
* https://fedoraproject.org/wiki/User:Adamwill/Draft_Fedora_21_test_plan
* https://lists.fedoraproject.org/pipermail/test/2014-May/121396.html
* https://lists.fedoraproject.org/pipermail/test/
Help Wanted (Soon) for Software Collection Package Reviews
----------------------------------------------------------
Software Collections is a way to package software into RPMs which sit
in `/opt`, outside of the main part of the distribution. The main use
case of Software Collections (SCLs, for short) is to have a stack of
software which moves at a different speed from the distro proper — in
RHEL or CentOS, that usually means faster, but in Fedora, it could be a
way to provide consistent runtime environments across multiple
releases.
Specifically, we’re planning to have a Ruby 1.9.3 with Rails 3.2.8 SCL
in Fedora 21 as a first experiment, led by Fedora contributor Marcela
Mašláňová. Since F21 will have Ruby 2.1.x with Rails 4.1, this will let
users with code that isn’t ready for the new version update to newer
Fedora — but port their own code on their own schedule
The Fedora Packaging Committee has been working on draft guidelines,
and the remaining big hurdle is that the 60-some new packages will need
to go through the standard review process. The plan is to make this as
painless as possible, by organizing an online activity day to crank
through them. If you’ve done some package reviews (and ideally if you
know Ruby), and are interested in helping, watch the Environments and
Stacks mailing list for more soon.
* https://www.softwarecollections.org/
* https://fedoraproject.org/wiki/Changes/Ruby193_in_SCL
* https://fedoraproject.org/wiki/User:Toshio/SCL_Guidelines_%28draft%29
* https://lists.fedoraproject.org/mailman/listinfo/env-and-stacks
--
Matthew Miller -- Fedora Project -- <mattdm(a)fedoraproject.org>
"Tepid change for the somewhat better!"
9 years, 11 months
uefi: What if ...?
by lee
Hi,
what if I got a new mainboard which uses uefi --- and might not even
support disabling or not using uefi --- and wanted to boot?
Booting is from a HP smart array P800. Would that work? Would I need
to do something to get it to work, and if so, what?
--
Fedora release 20 (Heisenbug)
9 years, 11 months
Problem with selinux and milter-greylist
by Will Yonker
Hi,
So I'm trying to get milter-greylist working with
selinux
and I seem to have a problem. It doesn't seem to know
what
milter-greylist is trying to access so I can't add a rule to fix
it.Â
Here is what I see in /var/log/message when I try to run
systemctl start
milter-greylist
May 27 12:47:07 dcsnow
setroubleshoot: SELinux
is preventing /usr/sbin/milter-greylist from
remove_name access on the
directory . For complete SELinux messages.
run sealert -l
f008afda-b837-4a7a-ad4e-80562d4ef31c
May 27
12:47:07 dcsnow python:
SELinux is preventing
/usr/sbin/milter-greylist from remove_name access on
the directory
.
*****Â Plugin catchall_labels (83.8
confidence)
suggests  *******************
If you
want to
allow milter-greylist to have remove_name access on theÂ
directory
Then you need to change the label on
$FIX_TARGET_PATH/>Do
# semanage fcontext -a -t FILE_TYPE
'$FIX_TARGET_PATH'/>where FILE_TYPE is one of the following:
greylist_milter_data_t,
var_run_t.
Then execute:
restorecon
-v '$FIX_TARGET_PATH'/>
*****Â Plugin catchall (17.1
confidence)
suggests  **************************
If you believe
that milter-greylist should be allowed remove_name
access on theÂ
directory by default.
Then you should report
this as a bug.
You
can generate a local policy module to allow
this access.
Do/>allow this access for now by executing:
#
grep milter-greylist
/var/log/audit/audit.log | audit2allow -M
mypol
# semodule -i
mypol.pp
In audit.log I
see:
type=AVC
msg=audit(1401209226.129:1909): avc:Â
denied { remove_name }
for pid=8467
comm="milter-greylist"
name="milter-greylist.sock" dev="sda6" ino=652403
scontext=system_u:system_r:greylist_milter_t:s0
tcontext=unconfined_u:object_r:var_t:s0 tclass=dir
Any ideas
on
how I go about finding out what needs to happen here?
Thanks in
advance for your help.
---
Will Y.
--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
9 years, 11 months