When running the transaction in a separate process it crashes if you use
a https repo source. There's really no need for threads or processes in
lorax so drop it.
Also switched to using the DNF TransactionProgress API for progress
reporting.
--
To view this pull request on github, visit https://github.com/rhinstaller/lorax/pull/50
The attached patch will allow a variant to provide a full url for its resources. Users of respin media or rebuilds (such as CentOS and Scientific) could utilize this to populate some community resources within the interactive install environment.
--
To view this pull request on github, visit https://github.com/rhinstaller/anaconda/pull/356
Missing /tmp/dd_disk file is not an error, changing to debug print instead.
*Resolves: rhbz#1251394*
--
To view this pull request on github, visit https://github.com/rhinstaller/anaconda/pull/349
Instead of defaulting to localhost.localdomain (which makes joining
an enterprise domain difficult), we will instead generate a random
hostname of the form <famous_person>.<adjective>.lan.
Note that if DHCP provides a hostname for this machine, the DHCP-
assigned hostname will be selected, rather than using this
randomly-assigned one.
Testing performed:
Booted two VMs, one bridged to a network which has DHCP-assigned
naming, one NATed with libvirt having no DHCP-assigned names. On
the bridged VM, anaconda selected new-host-4.home as directed by
DHCP. On the NATed VM, it selected newton.elegant.lan as directed
by this random name generation.
--
To view this pull request on github, visit https://github.com/rhinstaller/anaconda/pull/164
This is a request for comment, because it is incomplete. I am trying to figure out better ways to test the higher-level functionality but am having problems obtaining a cooperating virtual machine.
There are a few points:
* What makes this a bit tricky is that LUKS is handled as both a device and a format. There is a sense
in which it is both.
* I decided to start from the bottom up this time, working on resizing the format first, rather than worrying about the Actions.
* If the resize of a LUKSDevice is handled as just a resize of its slave's LUKS format then blivet.resizeDevice() does not need to change at all to handle LUKS devices AFAICT.
* However, it does not make much sense to resize a LUKS format outside the LUKS device...since the format knows nothing about the size of the, e.g., filesystem that it encrypts. And yet, it is still possible. Not sure whether it is necessary or desirable to do anything about that.
* I have not handled the dependencies between the actions.
--
To view this pull request on github, visit https://github.com/rhinstaller/blivet/pull/163
This enables Baytrail and similar atom CPUs that typically ship with a
32-bit firmware, but have a 64-bit capable CPU.
(This patch requires a newer shim package that is not ready yet, so don't pull this yet.)
Signed-off-by: Peter Jones <pjones(a)redhat.com>
--
To view this pull request on github, visit https://github.com/rhinstaller/lorax/pull/16