Red Hat Bugzilla – Bug 216677
install devours memory when the kickstart keyword repo is used
Last modified: 2008-03-05 19:05:08 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; fr; rv:1.8.1) Gecko/20060601 Firefox/2.0 (Ubuntu-edgy)
Description of problem:
when you create in the kickstart an entry with the new keywords 'repo' to use extras repositories, the install devours the memory eventualy leading to a crash.
added this in the ks.cfg I use, I have a server mirroring the fedora ones
repo --name=core --baseurl=http://yum/fedora/linux/core/6/i386/os/
repo --name=updates --baseurl=http://yum/fedora/linux/core/updates/6/i386/
repo --name=extras --baseurl=http://yum/fedora/linux/extras/6/i386/
repo --name=livna --baseurl=http://yum/livna/6/i386/
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1.configure repos in ks.cfg
2.start install, crashes at around 80-85% of install
crashes at around 80-85% of install
Please attach the error message you are receiving to this bug. Also, what
package selections are you making in the %packages section of your kickstart file?
just the install crashing red screen scrolling fast with rpmdb errors
rpmdb: PANIC: fatal region error detected; run recovery error: db4
Here's the %packages section
How much memory does your machine have?
The box where it happens has 1Gb RAM and I configured kickstart to create a 2Gb
SWAP partition on disk
I seem to be seeing this as well on a 512MB Core 2 Duo Latitude D620.
I've had successful installs with this same kickstart file on other machines
before now though. Maybe we've just finally grown big enough to cause problems...
I'm not sure it's a memory issue though. anaconda process size does grow for a
while when installing packages (as seen by top), to around 25%, but gets down to
about 3-5% again by the time of the crash. Unless that is really just showing
the anaconda process being slowing push to swap?
May just be a frail rpmdb (again)? I've been seeing problems on a box upgraded
Okay, free is now showing 141632k swap used, so memory is an issue.
Is this still happening with rawhide or a Fedora 7 test release?
(In reply to comment #8)
> Is this still happening with rawhide or a Fedora 7 test release?
Hard to test for a couple reasons:
- immediate but fixable: bug #231543 - multiple repositories are currently broken.
- harder - I didn't notice this problem until FC6 had been in production for a
while the the updates repo had gotten to an appreciable size. Not sure how to
similate this for testing.
I filed a new bug #241062, but I guess it's really a duplicate of this one, but
it points out the cause. rpms are building up in /tmp/cache/<repo>/packages and
filling up ram. This is why it showed up after fc6 had been released one enough
updates got released.