Anaconda-devel October 2011

anaconda-devel@lists.fedoraproject.org

25 participants
57 discussions

Start a nNew thread

[rhel5-branch] fix getMaxLVSize
by David Lehman 20 Oct '11

20 Oct '11

getMaxLVSize was using the lvm1 computation, but rhel5 uses lvm2.

1 2

[PATCH f16-branch] Copy all of live filesystem to target (#746844)
by Brian C. Lane 19 Oct '11

19 Oct '11

From: "Brian C. Lane" <bcl(a)redhat.com> Instead of depending on currentSize to be correct just copy the whole live filesystem over to the target. --- pyanaconda/livecd.py | 12 +++++++++--- 1 files changed, 9 insertions(+), 3 deletions(-) diff --git a/pyanaconda/livecd.py b/pyanaconda/livecd.py index 199f5c5..c55689f 100644 --- a/pyanaconda/livecd.py +++ b/pyanaconda/livecd.py @@ -152,7 +152,8 @@ class LiveCDCopyBackend(backend.AnacondaBackend): readamt = 1024 * 1024 * 8 # 8 megs at a time size = self.anaconda.storage.liveImage.format.currentSize * 1024 * 1024 copied = 0 - while copied < size: + done = False + while not done: try: buf = os.read(osfd, readamt) written = os.write(rootfd, buf) @@ -174,8 +175,13 @@ class LiveCDCopyBackend(backend.AnacondaBackend): copied = 0 continue - if (written < readamt) and (written < len(buf)): - raise RuntimeError, "error copying filesystem!" + if (written < readamt): + # Either something went wrong with the write + if (written < len(buf)): + raise RuntimeError, "error copying filesystem!" + else: + # Or we're done + done = True copied += written progress.set_fraction(pct = copied / float(size)) progress.processEvents() -- 1.7.6.4

2 1

Patches -- what else?
by David Lehman 19 Oct '11

19 Oct '11

Patch 1 is something old I've been meaning to fix. It is apparently not very often hit although it seems like it would be commonplace. It's only for master. Patch 2 is another, again harder, pass at fixing weird problems that only happen with the lvm dialogs. It's for f16-branch and master.

2 3

[f16-branch] fix autopart-with-shrink
by David Lehman 18 Oct '11

18 Oct '11

I broke this a couple of weeks ago. This time I tested "shrink existing system" to verify it works.

2 2

[rhel6-branch] mpath: flush more eagerly in filter_gui.
by Ales Kozumplik 18 Oct '11

18 Oct '11

Else we might end up with bad device names in exclusiveDisks. See also 861d48e0dcca256cd3c0c5cd72ca7ceed1848e09 and 8b8cf4cbee86e5bfb00802b5d5ab617f9c5d5fab. Resolves: rhbz#744772 --- iw/filter_gui.py | 14 ++++++-------- storage/devicelibs/mpath.py | 6 ++++++ 2 files changed, 12 insertions(+), 8 deletions(-) diff --git a/iw/filter_gui.py b/iw/filter_gui.py index 0c29c16..6da510d 100644 --- a/iw/filter_gui.py +++ b/iw/filter_gui.py @@ -629,15 +629,13 @@ class FilterWindow(InstallWindow): with open("/etc/multipath.conf", "w+") as mpath_cfg: mpath_cfg.write(cfg) - if anaconda.isKickstart: - # identifyMultipaths() uses 'multipath -d' to find mpath devices. In - # case this is an interactive kickstart install, it wouldn't display - # those already set up during the early storageInitialize() unless - # we flush them first. - log.info("filter_gui: flushing all multipath devices.") - iutil.execWithRedirect("multipath", ["-F"]) - (singlepaths, mpaths, partitions) = identifyMultipaths(disks) + # identifyMultipaths() uses 'multipath -d' and 'multipath -ll' to find + # mpath devices. In case there are devices already set up they won't + # show up (or show up with a wrong name). Flush those first. + log.info("filter_gui: flushing all multipath devices.") + flush_mpaths() + (singlepaths, mpaths, partitions) = identifyMultipaths(disks) # The device list could be really long, so we really only want to # iterate over it the bare minimum of times. Dividing this list up # now means fewer elements to iterate over later. diff --git a/storage/devicelibs/mpath.py b/storage/devicelibs/mpath.py index eb6e999..789add1 100644 --- a/storage/devicelibs/mpath.py +++ b/storage/devicelibs/mpath.py @@ -303,3 +303,9 @@ blacklist { ret += '}\n' return ret + +def flush_mpaths(): + iutil.execWithRedirect("multipath", ["-F"]) + check_output = iutil.execWithCapture("multipath", ["-ll"]).strip() + if check_output: + log.error("multipath: some devices could not be flushed") -- 1.7.6.4

2 1

[master] Do a better job of remembering if 'review and modify partitioning' was checked.
by Ales Kozumplik 18 Oct '11

18 Oct '11

The scheduling is reverted when going back thus it can not be used to remember user's choice that should preserve if reentering a particular screen. Resolves: rhbz#746703 --- pyanaconda/dispatch.py | 5 +++++ pyanaconda/iw/autopart_type.py | 6 +++++- tests/pyanaconda_test/dispatch_test.py | 12 ++++++++++++ 3 files changed, 22 insertions(+), 1 deletions(-) diff --git a/pyanaconda/dispatch.py b/pyanaconda/dispatch.py index 6aeae38..18bfa58 100644 --- a/pyanaconda/dispatch.py +++ b/pyanaconda/dispatch.py @@ -86,6 +86,8 @@ class Step(object): self.name = name self.target = target # None for dynamic target (e.g. gui view) self._sched = self.SCHED_UNSCHEDULED + self.client_data = {} # used by the client code, not reset when + # scheduling is reverted def _reschedule(self, to_sched, current_step): s_from = self.sched @@ -334,6 +336,9 @@ class Dispatcher(object): except errors.DispatchError as e: log.debug("dispatch: %s" % e) + def step_data(self, step): + return self.steps[step].client_data + def step_disabled(self, step): """ True if step is not yet scheduled to be run or will never be run (i.e. is skipped). diff --git a/pyanaconda/iw/autopart_type.py b/pyanaconda/iw/autopart_type.py index 536adcc..512d884 100644 --- a/pyanaconda/iw/autopart_type.py +++ b/pyanaconda/iw/autopart_type.py @@ -200,6 +200,7 @@ class PartitionTypeWindow(InstallWindow): self.storage.doAutoPart = True + self.dispatch.step_data("parttype")["review_checked"] = self.reviewButton.get_active() if self.reviewButton.get_active(): self.dispatch.request_steps("partition") # with kickstart bootloader is already scheduled to be skipped: @@ -253,7 +254,10 @@ class PartitionTypeWindow(InstallWindow): self.table = self.xml.get_widget("parttypeTable") self.prevrev = None - self.reviewButton.set_active(self.dispatch.step_enabled("partition")) + + step_data = self.dispatch.step_data("parttype") + self.reviewButton.set_active( + step_data.get("review_checked", self.dispatch.step_enabled("partition"))) self.encryptButton.set_active(self.storage.encryptedAutoPart) self.lvmButton.set_active(self.storage.lvmAutoPart) diff --git a/tests/pyanaconda_test/dispatch_test.py b/tests/pyanaconda_test/dispatch_test.py index b9243f6..9f2fac1 100644 --- a/tests/pyanaconda_test/dispatch_test.py +++ b/tests/pyanaconda_test/dispatch_test.py @@ -249,3 +249,15 @@ class DispatchTest(mock.TestCase): d.steps["betanag"].changes, {"betanag" : (Step.SCHED_SCHEDULED, Step.SCHED_UNSCHEDULED), "filtertype" : (Step.SCHED_SCHEDULED, Step.SCHED_UNSCHEDULED)}) + + def step_data_test(self): + from pyanaconda.dispatch import Step + d = self._getDispatcher() + self.assertEqual(d.step_data("filter"), {}) + d.step_data("filter")["key"] = 42 + # we can retrieve it + self.assertEqual(d.step_data("filter")["key"], 42) + # it persists over scheduling changes + d.schedule_steps("filter") + d._revert_scheduling("filter") + self.assertEqual(d.step_data("filter")["key"], 42) -- 1.7.6.4

2 1

[lorax] Add fpaste to install environment (#727842)
by Martin Gracik 17 Oct '11

17 Oct '11

--- share/ramdisk.ltmpl | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/share/ramdisk.ltmpl b/share/ramdisk.ltmpl index 221982e..b46ab43 100644 --- a/share/ramdisk.ltmpl +++ b/share/ramdisk.ltmpl @@ -70,6 +70,7 @@ install "fedora-gnome-theme" install "fedora-icon-theme" install "firstaidkit-engine" install "firstboot" +install "fpaste" install "ftp" install "gdb-gdbserver" install "gdk-pixbuf" -- 1.7.3.2

2 1

[Q] network configuration check
by Mykola Marzhan 13 Oct '11

13 Oct '11

Hi All, I am trying to add a network configuration check (patch is attached) to anaconda 13.21.117 (OL6.1). The check is added into "getNext" function on "network" page in gui mode. But it is working only if network is configured before (for example in "updates" mode). I can't find a solution how to resolve this issue. Do you have any ideas? -- Mykola Marzhan PortaOne, Inc., Release Engineer delgod(a)portaone.com Tel: +1-866-SIP VOIP (+1 866 747 86 47) ext. 7159

1 0

Potential F16 patches
by David Lehman 12 Oct '11

12 Oct '11

These are both for f16-branch and master. The first one is sort of a best-effort at forcing boot device selection even for custom partitioning. This leaves only kickstart without forced boot drive selection. Obviously we'll want to do this better in the new UI. The second one is the beginnings of handling for systems that have shown to be incapable of booting from a GPT-labeled disk. If we want to make a blacklist at some point, fine. Here's this for now.

3 4

[PATCH rhel5-branch] Translate network interface configuration GUI dialog (#709931)
by Radek Vykydal 12 Oct '11

12 Oct '11

Resolves: rhbz#709931 --- iw/network_gui.py | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/iw/network_gui.py b/iw/network_gui.py index fcd315f..2549249 100644 --- a/iw/network_gui.py +++ b/iw/network_gui.py @@ -543,7 +543,8 @@ class NetworkWindow(InstallWindow): class NetworkDeviceEditWindow: def __init__(self, netwin): self.netwin = netwin - self.xml = gtk.glade.XML(gui.findGladeFile('netpostconfig.glade')) + self.xml = gtk.glade.XML(gui.findGladeFile('netpostconfig.glade'), + domain="anaconda") # Pull in a ton of widgets. self.toplevel = self.xml.get_widget("net_post_config_win") -- 1.7.4

2 1

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Anaconda-devel October 2011