Bug 594803

Could you attach /tmp/storage.log and the output of udevadm info --export-db (switch to tty2, hit up arrow a bunch) to this bug report?

Created attachment 416192 [details]
storage log

Here's the anaconda storage.log from the installed system. I'll try and get the udevadm output once I'm in a position to do the test install again.

According to your storage.log, you have two devices:  sda is a USB device of some description, and sdb is a drive with two partitions on it.  Are you doing a hard drive installation method where a USB device is the source?

sda is a USB card reader. It had no media in it, and was not ever shown in the UI.

Does updates=http://clumens.fedorapeople.org/594803.img fix this issue and not drastically slow everything down?

It doesn't slow anything down (that I could notice), but it doesn't fix the issue either.

Created attachment 416993 [details]
storage log w/updates image

storage log w/updates image

Created attachment 416994 [details]
anaconda log

Can you also jump over to tty2, start python, and paste the output of:

> import parted
> partedDevice = parted.Device("/dev/sda")
> print int(partedDevice.getSize())

> partedDevice = parted.Device("/dev/sda")

_ped.IOException: Error opening /dev/sda: No medium found

Looks like the populate() step that does the skipping for the device is done after we've decided to show the menu b/c there are there are two devices total.

Right, that's what I was trying to solve with my updates.img:

diff --git a/iw/filter_gui.py b/iw/filter_gui.py
index a85cd1f..9f5b2f5 100644
--- a/iw/filter_gui.py
+++ b/iw/filter_gui.py
@@ -520,6 +520,8 @@ class FilterWindow(InstallWindow):
             self.buttonBox.show_all()
 
     def getScreen(self, anaconda):
+        global totalDevices
+
         (self.xml, self.vbox) = gui.getGladeWidget("filter.glade", "vbox")
         self.buttonBox = self.xml.get_widget("buttonBox")
         self.notebook = self.xml.get_widget("notebook")
@@ -575,12 +577,14 @@ class FilterWindow(InstallWindow):
         (raids, nonraids) = self.split_list(lambda d: isRAID(d) and not isCCISS(d),
                                             singlepaths)
 
+        self.populate(nonraids, mpaths, raids)
+
         if anaconda.id.simpleFilter:
             # In the typical use case, the user likely only has one drive and
             # there's no point showing either the filtering UI or the
             # cleardisks UI.  Unfortunately, that means we need to duplicate
             # some of the getNext method.
-            if len(singlepaths) == 1:
+            if totalDevices == 1:
                 anaconda.id.storage.exclusiveDisks = [udev_device_get_name(singlepaths[0])]
                 return None
 
@@ -592,8 +596,6 @@ class FilterWindow(InstallWindow):
                           self._makeMPath(), self._makeOther(),
                           self._makeSearch()]
 
-        self.populate(nonraids, mpaths, raids)
-
         # If the "Add Advanced" button is ever clicked, we need to have a list
         # of what devices previously existed so we know what's new.  Then we
         # can just add the new devices to the UI.  This is going to be slow,

This request was evaluated by Red Hat Product Management for inclusion in a Red
Hat Enterprise Linux major release.  Product Management has requested further
review of this request by Red Hat Engineering, for potential inclusion in a Red
Hat Enterprise Linux Major release.  This request is not yet committed for
inclusion.

I think I'm going to need access to a machine with this problem, as I don't see anything else obvious in the code that could be a problem.

Created attachment 417795 [details]
patch

The issue with the prior patch was that populate() was called before the pages were set up, so totalDevices would never get incremented correctly. On fixing that, I noticed that there would still be an issue because simplepaths could still contain multiple devices.

The attached fixes it by filtering simplepaths based on the size of the device. I decided to read it out of sysfs, but we could create a partedDevice here and do it too... just seemed overkill to init parted on each device.

Is that patch on top of the one I initially put into this report?

It's instead of that patch.

I have an ASUS eeePC laptop with integrated USB card reader to test this bug. However I've hit bug #612232 while trying snapshot #7.

*** Bug 558689 has been marked as a duplicate of this bug. ***

This is not fixed with snapshot #7 / anaconda 13.21.56. 

I have a single disk (sdc) and USB SD card reader (sdb) with no media in it. The Basic storage screen asked me which disk I want to install to and was showing only sdc. Moving back to ASSIGNED.

Will attach storage.log shortly.

Also I went to tty2 and executed:
> import parted
> partedDevice = parted.Device("/dev/sdb")
 a traceback follows: _ped.IOException: Error opening /dev/sdb: No medium found

which is expected.

Created attachment 431847 [details]
storage.log

Created attachment 431848 [details]
tarball of all logs from stage2