Bug 523567 - Install is impossible using harddrive method with less that 256Meg of ram
Summary: Install is impossible using harddrive method with less that 256Meg of ram
Keywords:
Status: CLOSED CANTFIX
Alias: None
Product: Fedora
Classification: Fedora
Component: anaconda
Version: rawhide
Hardware: All
OS: Linux
low
medium
Target Milestone: ---
Assignee: Anaconda Maintenance Team
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2009-09-16 01:32 UTC by Jerry Vonau
Modified: 2009-09-17 15:45 UTC (History)
2 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2009-09-16 15:51:21 UTC


Attachments (Terms of Use)
patch (729 bytes, patch)
2009-09-16 01:39 UTC, Jerry Vonau
no flags Details | Diff
patch2 (2.69 KB, patch)
2009-09-17 15:45 UTC, Jerry Vonau
no flags Details | Diff

Description Jerry Vonau 2009-09-16 01:32:24 UTC
Description of problem:

https://fedoraproject.org/wiki/Documentation_Beats_Hardware_Overview
Recommended for text-mode: 233 MHz G3 or better, 128MiB RAM. 

This is impossible when anaconda's loader is copying install.img 111M into /tmp plus unpacking the initrd of 22M...     

Version-Release number of selected component (if applicable):
Sept 15 boot.iso

How reproducible:
boot the install from usb

Steps to Reproduce:
1. download boot.iso
2. prepare usbdisk
3. edit the boot loader's config file appending :/images/install.img to stage2=hd:LABEL=Fedora 
This is required otherwise loader falls back to asking for path to install.img
This could be fixed by changing the boot line on the boot.iso, I think it would be the same anyway.
4. install fails mounting /tmp/install.img looking at tty3
  
Actual results:
can't boot the install on my XO. 

Expected results:
A little farther into the install at least ;)

Additional info:
If I spin up my own loader using attached diff I can at least proceed until I crash when my unsupported network causes issues.

Comment 1 Jerry Vonau 2009-09-16 01:39:05 UTC
Created attachment 361170 [details]
patch

Comment 2 Chris Lumens 2009-09-16 15:51:21 UTC
Unfortunately, we have to copy the image over in the preupgrade case at least (see commit 536fc1564ae4506c1ba684d0380fc0f665d4731d) so we can later unmount the partition holding the image.  Otherwise, there's tracebacks when we attempt to mount/unmount things already mounted.  It's likely we will hit this problem in other places too.

Comment 3 Jerry Vonau 2009-09-17 15:43:33 UTC
# This takes care of mounting /mnt/isodir first.
self._switchImage(1) is only called if preupgrade is absent, and then mountDirectory is used, that checks if /mnt/isodir is already mounted.

What if loader was changed to really use the preupgrade flag to tell the difference, would that make you a little less resistive to the change?

Patch to be added.

Comment 4 Jerry Vonau 2009-09-17 15:45:21 UTC
Created attachment 361519 [details]
patch2


Note You need to log in before you can comment on or make changes to this bug.