[PATCH rhel7-branch] Process excludedGroupList after packages and groups.
by Chris Lumens
The order we had things in was causing the excluded group to get
removed, and then packages to get installed afterwards. This is clearly
the wrong order and had the effect of the conflicts groups not working
like they are supposed to.
Resolves: rhbz#1249036
---
pyanaconda/packaging/yumpayload.py | 12 ++++++------
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/pyanaconda/packaging/yumpayload.py b/pyanaconda/packaging/yumpayload.py
index a21ad16..ff7ef14 100644
--- a/pyanaconda/packaging/yumpayload.py
+++ b/pyanaconda/packaging/yumpayload.py
@@ -1308,12 +1308,6 @@ reposdir=%s
except NoSuchGroup as e:
self._handleMissing(e)
- for group in self.data.packages.excludedGroupList:
- try:
- self._deselectYumGroup(group.name)
- except NoSuchGroup as e:
- self._handleMissing(e)
-
for package in self.data.packages.packageList:
try:
self._selectYumPackage(package)
@@ -1323,6 +1317,12 @@ reposdir=%s
for package in self.data.packages.excludedList:
self._deselectYumPackage(package)
+ for group in self.data.packages.excludedGroupList:
+ try:
+ self._deselectYumGroup(group.name)
+ except NoSuchGroup as e:
+ self._handleMissing(e)
+
self._select_kernel_package()
self.selectRequiredPackages()
--
2.4.3
8 years, 8 months
Updated: [rhinstaller/anaconda/pulls/314 master] Add test documentation
by dashea
This is the start of what is supposed to be a human readable test documentation of the Anaconda test suite. Filing here for comments.
I'd like to also include API docs for the Python files under tests/ but so far I'm having troubles building these. I will likely have to add __init__.py files to all directories that are missing them and fix the import failures from Sphinx (e.g. by altering PYTHONPATH if necessary).
Btw it looks like the generated HTML is under version control, not sure if that needs to be removed.
--
To view this pull request on github, visit https://github.com/rhinstaller/anaconda/pull/314
8 years, 8 months
New: [rhinstaller/blivet/pulls/215 master] Do not reserve extra space for metadata in a VG with RAID PVs
by dashea
Reserving 5 extents per disk per LV for metadata if a VG is on top of RAID PVs
is a weird magic that saved us troubles with placing LVM on top of RAID, but
even on the first it's just wrong and doesn't make sense. LVM doesn't allocate
any extra meta data if an LV is added to a VG (unless it is a meta data LV of
course) and the RAID device below it (the PV) is just a block device with some
size and no matter how many disks are below, the meta data space used by LVM is
just subtracted from the RAID device's (PV's) size. Moreover the LVM extent size
has nothing to do with the RAID device so it's like mixing apples and oranges.
If things don't fit we need to tweak our calculations of free space/size in/of
an MD RAID device (depending on the level and underlying disks) since that's the
only place where these tweaks really make sense. However, after quite a lot of
testing it seems that we are reserving more than enough space for the meta data.
--
To view this pull request on github, visit https://github.com/rhinstaller/blivet/pull/215
8 years, 8 months