Red Hat Bugzilla – Bug 444921
preupgrade doesn't handle /boot being out of space
Last modified: 2008-05-02 12:33:11 EDT
Description of problem:
preupgrade fails to compute required space on /boot. Should there be
insufficient space, it fails to download the image (in my case stage2) and
tosses up an error, but still changes grub and prompts the user to reboot as if
it were successful.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. make a small /boot (of about 100MB) on a f8 system
2. run preupgrade
preupgrade fails to download stage2 image file (minstg2.img ~60MB) due to space
constraints but still changes grub.conf and prompts user to reboot.
preupgrade would fail to run due to lack of space on /boot, or at least not
Doesn't anaconda's "autopartition my disk" feature default to 100MB /boot?
If you *do* have insufficient space in /boot, we just download stage2 after you
reboot instead. The error message now says this pretty clearly:
"Not enough space in /boot to cache stage2.img.
It will be downloaded once the installer starts."
Once you reboot it fetches stage2.img and the install continues normally.
This only happens if you have an older system or you intentionally create a
small /boot. F8 and later default to 200MB /boot for exactly this purpose.
Yeah, I kinda figured out the 'it grabs stage2 later' after a few less than
useful attempts at the reboot and looking at the grub entries. The updated
error message should be enough to mollify the ultra-stupid such as myself, thanks!
Presently I'm getting at error at reboot about my groups info being wrong due to
a bad composition regardless of what mirror I try or blowing away all the stuff
in /var/cache/yum/anaconda-upgrade. skvidal looks to have built an 0.9.3 this
morning, so I'll try that and open another bug if I have trouble or just hop on
Or I could just ramble away in this ticket forever....
See bug 444915: Unable to read group information with preupgrade