| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
| |
This identifies that a device is part of a multipath, and builds an
mpath device for partitioning.
|
| |
|
|
|
|
|
|
|
|
|
| |
During testing I noticed that anaconda did not recognize pre-existing
logical volumes and thus that I could not upgrade / rescue an existing
default partitioning install.
I tracked this down to the last lvm call in 70-anaconda.rules failing
(udevadm test is your friend). Passing --ignorelockingfailure to lvm
fixes this. For good measure this patch does this for all invocations in
70-anaconda.rules
|
| |
|
|
|
|
|
| |
We are currently not calling blkid when a block device is not an
mdraid set or a device mapper device, resulting in us not identifying
mdraid members as such. This probably worked when still using vol_id
because some other rules file was doing an import vol_id before our rules.
|
| |
|
|
|
|
| |
udev no longer ships a vol_id binary, so we have to use blkid
instead. Require a new enough util-linux-ng that this does what
we want
|
| | |
|
| |
|
|
| |
This is needed for live installs since they run on a normal system.
|
| |
|
|
|
| |
This will enable us to detect filesystems or other formatting
on md arrays.
|
| |
|
|
|
| |
In the anacond udev rules do not execute the 2 import commands which depend
up on the vg_name if we don't have a vg_name.
|
| |
|
|
|
|
|
| |
The way the anaconda udev rules are currently written, they require
the device nodes to already be present when they run. This is not always
the case (as udev creates the nodes after running the rules). This patch
fixes this, which fixes the recognition of lvm VG's and LV's on iscsi disks.
|
| | |
|
| |
|
